Podcast Transcript: Mosen At Large 105, the AccessiBe controversy. Can AI make the web fully accessible in a few short years, or might it make matters worse?
You can read the transcript below, Download this episode in Microsoft Word, or Download this episode in accessible PDF
Jonathan Mosen: In episode 105, can artificial intelligence make inaccessible websites a thing of the past, or are they high-tech snake oil that can cause more problems than they solve? The entire show today is devoted to accessibility overlays and their current leading proponents, AccessiBe.
Jonathan: Thank you for listening today. I appreciate it. If you would like to be in touch with the show and you may well want to respond to some of what you’re hearing today, let’s go through the contact information. My email address is firstname.lastname@example.org and that’s J-O-N-A-T-H-A-N@mushroomfm.com. You can send in an email attachment recorded on your PC, your smartphone, anything that’ll record an MP3 or an M4A file and attach it to the email. Or you can just write the email down if you prefer. The listener line is open as well. That number is in the United States. You can record a voicemail message there at 864-60MOSEN. That’s 864-606-6736. Particularly, if you were listening to the show live on Mushroom FM or the Mosen At Large Facebook page or YouTube channel, you can use the Twitter hashtag #mushroomfm. Just write the number sign and Mushroom FM all joined together.
Listeners can also, of course, track that hashtag. Just search for the #mushroomfm, keep that search open and you’ll be able to see what other people are tweeting about and respond. We can have a bit of a conversation as our guests explain their positions on what we’re going to be talking about today. You can also comment on our Facebook page and YouTube channel as the audio plays out. I will do my best to monitor that as well and approve those comments as they come in.
This is a unique show because it’s all about one thing and it’s an important thing. One of the biggest barriers that we face as blind people is access to information of all kinds. The first step is making sure that we get access to the right assistive technology to meet our particular needs and then we know how to use it.
Some of us can pick that up ourselves, others need a helping hand to do that. Then, if we can learn to use the world wide web well, it can be a game-changer. Those of us who can remember a time before the web know how much easier it is now to just browse leisurely at a store or transact personal business. Many of us of a certain age will remember going into a bank and the bank teller yells at the top of their voice what your bank balance is because, of course, you’re blind so you can’t hear them.
Reading the newspaper, of course, is now easier than it’s ever been. Just generally taking part in the public discourse is much more possible thanks to the web. But, and there’s always a but, isn’t there? As you know, there’s a major caveat. Websites must be accessible through a combination of things like awareness-raising and a little bit of legal action here and there. People like to think that the web has become a much more accessible, inclusive place but there is much left to do.
We have all experienced the frustration of going to a website only to find that its developer has not followed the simple steps required to make their site accessible. In classic capitalism fashion, where there’s a problem, someone sees an opportunity, and entrepreneurs have seen the opportunity. Several companies have created technology they believe can make accessibility easier to implement for businesses, making the web a much more accessible place for us all.
Some say this technology is revolutionary and in a relatively short time could make inaccessible websites a thing of the past. Something they say the current traditional approach will never do. Others say we’re seeing a looming accessibility crisis as businesses purchase what is nothing but high-tech snake oil thinking erroneously that they’ve solved all their accessibility problems and inoculated themselves against lawsuits when in fact, they may have made matters worse.
The most talked-about company in this space is AccessiBe. It would appear to be the most deployed accessibility overlay. It is very well-funded and they’ve recently appointed a high-profile name in our community, Michael Hingson, as their chief vision officer. Now, if you’ve visited a website lately and possibly heard a sound, being told to press ALT+1 for screenreader mode, you’ve pressed that key or pushed that button and you’ve noticed some reformatting of the web page, and perhaps a few additional image descriptions that have been generated automatically. Then chances are you have used accessiBe’s technology.
Tensions and passions are high on social media and in the accessibility community about accessibility overlays in general and AccessiBe in particular. That’s not surprising because, given what’s at stake, it’s one of the most important issues we face as a blind community. That’s why the entire three hours of the show are devoted to these subjects.
This week it’s my goal to give you a range of perspectives and to question those who have strong views on either side so you can make up your own mind. There is a lot of information here. If you’re listening to the podcast version of the show, you’ll be pleased to know it’s segmented. We always make extensive use of chapter marks on this show.
If you’re listening in a podcast client that supports chapter marks, and most do, you’ll be able to skip forward and back through the different interviews right from within your podcast client, but if you don’t have such a client, I have also provided extensive timestamp information in the show notes.
Sam Evans is a well-connected and respected member of the web accessibility community and I thought we would talk with her about her perspective on AccessiBe and accessibility overlays in general and whether this technology has any promise. Welcome back, Sam. Nice to have you back on the podcast.
Sam Evans: Jonathan, thanks so much for contacting me. I’m so pleased to join you once again.
Jonathan: Now, we’ll have listeners who are very familiar with this whole issue. If you’re on social media and following certain people, it’s pretty difficult to miss, but then we will also have listeners who are new to this whole discussion. I thought you would do a great job for us of explaining what these accessibility overlays are, what they are seeking to achieve.
Sam: From my perspective and I’ll premise this by saying that I work with a community of disabled people, blind people, accessibility professionals, and practitioners whose sole effort, in both their personal and professional lives, is to make the digital experience inclusive for everyone and accessible for everyone using their technology of choice.
Accessibility overlays have become really popular. I think that they’re part of a wave of interest and business opportunities, probably from an entrepreneurial perspective of, “Hey, people are concerned about this. They don’t want to be sued from the US perspective or taken to the ombudsman in the European perspective. We’re going to create one line of code that will fix it all.”
It sounds brilliant but the challenge is accessibility in an artificial intelligence perspective, from AI, the computer doing it for you can’t really fix it. I think there is a variety of types of overlays that we see now related to websites and they’re used different ways. Some are presented by accessibility companies that actually work in accessibility that work towards digital inclusion and making the actual experience usable by the person who’s using assistive technologies.
Platforms where you go to a company and they create a website for you, but you can’t actually get to the underlying code to remedy those problems that make the websites inaccessible. There are accessibility companies that write plugins that allow you to overwrite that without diving into the code. You’re not breaking anything to actually implement accessibility best practices.
Then there are accessibility overlay companies that I don’t believe are actually accessibility companies because they present to the public, “Just do this. You don’t have to know anything about accessibility and it’s going to fix everything for you.” But if you actually read their terms of service, they won’t actually offer what they market. They’ll promise to make your site accessible but many of them actually say, “You won’t be accessible unless you were already accessible before you installed our product,” which is a bit of a circular logic and a bit defeatist for their supposed intent.
Many of them introduce additional problems and challenges for people who use assistive technologies. Whether it’s a selective color-coding that looks a bit like– If people are using reading technologies that use their eyes for vision, they might highlight a line of code or line of texts that you would see on the screen and it scrolls up and down but that’s very distracting for many other people. A lot of the overlays don’t work with screen readers. They don’t work with text to speech and they don’t work with input other than a mouse. Which then begs the question, what are they solving?
Jonathan: When I’ve been looking at this as a user, I see a number of things going on. Sometimes you see that when you turn this mode on, you might get a few more regions on the page. Sometimes there might be text that’s in a larger font that isn’t rendered as a proper HTML heading for your screen reader to navigate to with this off, but AccessiBe is seeing it as a heading and rendering it as such when you turn it on. Do you think there’s ever a situation where they might add value, or is it just fundamentally flawed technology at this point?
Sam: I am a bit biased because I don’t think that you can simultaneously promote in advance inclusion by monetizing barriers and telling people they don’t need to learn how to do it well or for the right purposes. I’ve not seen the opportunity where this third party of overlays advances accessibility. They introduce things that are visual in nature, which we know many people who use screen readers do not have vision or don’t use vision to navigate their digital technologies. Presenting something much like style headings in CMS products just because it looks like it’s a big blue letter doesn’t make it a header that anyone can navigate.
I don’t find the value add in them. I find a big danger in the damage that’s done in the work towards inclusion and accessible content. It wouldn’t take any more programmers to correct these websites than it does to program them incorrectly to start. What happens is there are companies and organizations that think, “Everything is fixed. It’s all great. I have nothing to worry about,” but they aren’t going to discover until they lose those customers. Their customers can’t pay their rent. They can’t pay their bills. They can’t find the content they need.
The other thing these third parties, this third group of overlays does, is they assert that so-so alternative text or a text description, or a text equivalent of a graphic. It doesn’t really matter. One that I saw last week is supposed to be for a high-end retail shop, a high-end retail clothier. The text said, “Woman folding laundry,” but the image was of a very flashy-looking woman standing, picking through very flashy-looking clothes as if she were having a delightful shopping experience. These products don’t validate because they can’t. They can’t say this is relevant information.
They don’t tell you that it’s a relevant text link, so you’re still not going to know if you don’t have vision what the link is taking you to. They also assert that graphs and data don’t exist on websites, so they don’t address it. I’m trying to imagine how this concept exists in the business world to say that data and tables and infographics don’t exist in the 21st century.
Jonathan: On the flip side, on their website, AccessiBe is arguing that expecting businesses, many of which are small operators, the majority of businesses are one or two-person operations, expecting them to pay an accessibility consultant megabucks isn’t working. That’s why there are so many websites that are still difficult for assistive technology users to access. They say that with their technology, they can get the web fully accessible by 2025. They say that when you look at the number of inaccessible websites out there, the manual approach clearly has not worked.
Would we be better off as a community then, if we worked with accessibility overlay companies so that they work better and more of the web can be made accessible? Is there some way this can be made to work?
Sam: I think the challenge is that many of the tenets of web accessibility and the best practices in web accessibility are tied to the tenets of inclusive design, not just for websites, but for documents, for design, for other elements. If we dismiss that as the core, it’s not any different than a built environment saying, “In order to have an accessible structure, you must make this a usable experience from the front to the step to the doors, in all of the rooms avoiding the reality of making inclusive design.” The best practice is not making progress, so let’s talk to what their claim is.
They claim that small shops would have to pay $10,000 for an accessibility remediation. The reality is these websites that are set up for $150 or $200 are mom-and-pop or “it’s my side gig.” We helped a really big one of those take an overlay off their website. It took me, not a technical expert, two and a half hours to explain to them how to choose an accessible WordPress theme and have everything looking back the exact same way. Two and a half hours, you’d be maxed out at about $500 to fix that if you paid somebody that was a high-end accessibility expert.
Jonathan: This is what got me curious when I was doing my research for this show today. There are so many accessible themes and templates out there for the major CMS. I apologize, I didn’t dejargonize the CMS thing. The content management systems like Drupal and WordPress, which are the two big ones, and WordPress is by far the biggest, but there are many accessible options that actually adhere to good coding, so it’s not necessarily an expense, is it? It’s a public education thing about making sure you make the right choices.
Sam: I would agree. The last time I looked, which is a few months ago, WordPress had 97 accessible themes. I just helped this agency change a theme. Literally, it was, “Put it in the sandbox, change the theme. Let’s make sure it works. Here you go. Do these five things with your content, and you will be good.” They were done. In the majority of websites, the biggest challenges are things like color contrast, contextual links, alt text, simple things that you don’t have to have an advanced technical background to work with if your underlying core is done.
Picking the right content management system theme or having somebody come in and change your CSS, if you don’t know what that is, your content style sheets, if they can get to that and make edit, we are talking about small changes, and we are not talking about thousands of dollars for the average small business. We’re not. It just simply isn’t true. It’s a scare tactic. When people use scare tactics to scare people away from talking about disability and inclusion, they’re working quite counter to the premise of accessibility and inclusion. That’s dangerous, in my opinion.
I think that the premise that accessibility in any format is a checklist of one and done or one line of code and you don’t have to worry about it is dismissing disabled people across the board and suggesting that I don’t have to worry about those people anymore as opposed to, “I want the human experience to be as robust and inclusive as possible, so I’m willing to learn how to do this in a better way.”
Jonathan: You have also been quite vocal on Twitter, in particular, about the way that AccessiBe is marketed. Can you tell me about some of the concerns you have regarding representation of the product and the service?
Sam: If we are honest about what we’re selling and what it does and doesn’t do, I wouldn’t have an issue, but what they seem to advertise is not at all what they do. It doesn’t actually make a website accessible, usable, or compliant. Their own terms of agreement expressly say that they will not do that, and it doesn’t protect you from litigation. They’re happy to state it if you read the [unintelligible 00:17:53] or find the right policy page. Their premise is and it’s a brilliant marketing campaign, and they’ve obviously put their venture capital into marketing and advertorials, but what they’re selling is not what they’re actually delivering in product. It’s a bit disingenuous.
Jonathan: AccessiBe has said on their blog that they have been subjected to a disproportionate amount of harassment. They’ve also described what they perceive as hostility directed towards them. Sometimes social media beat-ups can get out of control. Has the quantity and the tone of the dialogue been proportionate on social media about AccessiBe?
Sam: What I’ve seen for more than a year is disabled people, blind people, deafblind people, people with reading disabilities and cognitive disabilities, give them direct, specific, actionable feedback and they deny every piece of feedback that’s provided to them but then ask people to come off to the side in quiet, not in public, and want people with disabilities to do usability testing for them, which is a product and service they should be paying for on their own.
They have gone through and claim they have addressed this, but several members of their staff and employees have come out and been beyond aggressive and calling disabled people and accessibility practitioners and professionals grifters and greedy consultants and trying to take tens of thousands of dollars from poor businesses. That was very much one of their tactics for the better part of the year. They have not really engaged with the disability community or the accessibility community. They’ve been invited to participate, but they have not yet done that because and rob their business. They are certainly allowed to control their engagement, but if their premise is, “We support accessibility and inclusion,” they aren’t really taking that and doing anything with the feedback. They keep asking for not generalizations and people are giving them specifics. If you were to do it the right way in the business world and pay for usability and accessibility testing, that’s a pretty pricey thing, but they want it all for free. Supposedly, if what they claim is true that they are an accessibility company, they would be doing this usability testing on their own, they would have encountered these challenges, they would know the blocks.
They would know that the overlay blocks content when you use screen readers, they would know that that’s not really a heading that can be navigated, but they’re not willing to have those discussions in public.
Jonathan: Some people would say, you now say I’m a representing the establishment, and it’s a bit like how we’ve seen disrupters come along before, whether it be the music industry and the streaming services that we have access to now, or the movie industry, the motor car when everybody was using horses. There are times that technology comes along that upends everything, and it’s usually the establishment that’s being upended that is the most resistant.
Is there a danger that that is what we are seeing? That this is actually long-term, a good thing.
Sam: Anything is possible, however, it has taken 30 years of people working together to get some baseline content together because accessibility isn’t just about websites, it carries over those core tenets crossover into other areas of our world and our lives and learning and the education and work. Dismissing the need for things to be coded properly and I know it will take time.
Yes, there are disruptors, but invested some of that venture capital into education and programming so that programmers could learn to do things the right way in tandem with their product, I might be more apt to think they were actually part of the community, but they’re not working on any of that.
Jonathan: I see AccessiBe is saying now that they are going to join the World Wide Web Consortium, and that’s where the web accessibility guidelines come from. What do you see as the long-term answer here? Can AccessiBe and the accessibility industry coexist? Or would you rather they just went away?
Sam: If they would be honest about what they do and what they don’t do, they can play in the market as long as they’re able.
Jonathan: Like the show, then why not like it on Facebook too? Get upcoming show announcements, useful links, and a bit of conversation. Head on over now to facebook.com/mosenatlarge. That’s facebook.com/M-O-S-E-Natlarge to stay connected between episodes.
Male Speaker: Jonathan Mosen, Mosen At Large podcast.
Jonathan: If you have any comments you’d like to share with us on this topic, you can drop me an email to Jonathan, that’s J-O-N-A-T-H-A-N@mushroomfm.com. That email can contain an audio clip if you would like using your smartphone or your computer. Or you can write the email down and I will read it. You can also call our listener line, bear in mind, given the complexity of the subject, there is a five-minute limit on that voicemail. 864-60-MOSEN is the number, that is a US number 864-606-6736.
I’m going to read a listener email now from Steven Clower, it is lengthy and it also holds some strong opinions. Rest assured that I will be putting many of these to Michael Hinkson later when he comes on the show from AccessiBe to talk about the product and some of the concerns that have been expressed. Steve says, “I wanted to pass along my experiences with the AccessiBe overlay, and I hope, convey just how dangerous this type of technology can be for those of us reliant on-screen access software.
To start with, I have been fighting for access in my personal life and on behalf of others for over 20 years now. I put up with high school instructors who refused to teach blind students. I had to write custom programs on a braille Lite to emulate statistical functions for a college-level course I took in my senior year of high school, which my peers got for free on their handy-dandy TI-83 calculators, and went back and forth for months to get permission to use my meager creations with the college board who oversees advanced placement exams as well as the SAT in the United States.
I performed similar work throughout college. More recently, I worked on projects such as implementing the initial version of the Health and Human Services office on disability home page, supporting and writing features for the now-retired Window-eyes screenreader wrote the GWconnect Skype client, which was popular some years back, performed accessibility remediation services for states ACA websites.
Added screen reader access to the ham exam online testing system so prospective or existing blind radio amateurs can take remote exams in 2020. I have most recently helped make one of the most popular online math tools, Desmos, accessible to persons with disabilities. I mentioned all this to establish that I am conversant with accessibility standards across a wide range of environments and have plenty of experience implementing them in the real world.
My first encounter with the AccessiBe overlay occurred last summer. I use Namecheap, www.namecheap.com as my domain registrar and web host. The time had come to make some changes to my account. I forget specifically what I was there to accomplish or renew. Regardless, I do remember very clearly a button instructing me to press it for something called screenreader mode.
I also noticed the page was alerting me to press the button precisely every 10 seconds. I loathe the separate but equal philosophy as it never pans out but as standard interaction with the page was causing my focus to jump all over the place. After being continuously harassed by the website, I decided to give it a try. Boy, was that ever a mistake. The result of clicking the button was a single lengthy block of text, which comprised the entire inner text of each webpage item.
There was absolutely no actionable interface elements, including a way to disable this new, horrible mode in which I found myself. What was worse was that I found a cookie stored in my browser that forced screen reader mode automatically when I reloaded the page, including all browser restarts. The solution at the time was to clear out all my browser cookies since I wasn’t aware of the responsible vendor at the time.
Luckily, on a desktop browser, this worked. I was treated to an equally unusable experience on my iPhone but had a much more difficult time removing the cookie. I did eventually manage to accomplish what I required. Shortly after that, I sent Namecheap a message, letting them know that their screen reader mode didn’t work. They responded with a generic message to the effect of ‘Thanks for the feedback.’
Today, the same overlay is still present and the site equally unusable. If you dare to try screen reader mode, be sure you have the ability to clear your browser cookies first. I next encounter their overlay when I was trying to pay rent a few weeks later. Our apartment’s bill processor at the time had extremely poor accessibility. Before that day, I had been able to work with it well enough to take care of the essentials.
When I loaded the site on that fateful morning, the first thing I heard was press ALT+1 for screenreader mode. This sure felt familiar, especially, once I began to hear the prompt repeated every 10 seconds. Curious, I decided to give it a try. The results weren’t quite as bad as Namecheap. I was at least still given the option to turn the screen reader mode off by pressing ALT+0.
Most distressing of all, however, was that the workflow for adding a bank account and scheduling a payment had been effectively neutered. No matter what I tried, I could not make the interface respond. Even when switching back to the “cited version” which also brought the return of those continuous ARIA alerts. I eventually wound up calling a sighted friend to help me write a check, something I have really had to do, how humiliating and what a tremendous step backwards.
After visiting their home page, I knew I had hit the answer. AccessiBe, a company claiming to solve so many accessibility problems for us poor disabled folk was responsible for these disastrous user experience regressions. I also found that others had already picked apart their software, pointed out obvious problems, submitted them to AccessiBe, and was summarily ignored or labeled opportunist. Needless to say that I knew I would get nowhere trying to talk to AccessiBe, after all, they had just swindled some venture capitalists out of millions of dollars. I just want to insert a narrator’s note here and say that is obviously a listener opinion. Returning to the text, as they say, their software had nine flaws that to this day continue to pervade the web. Their most recent attempts to silence, threaten and dismiss critics affirm that I was right to not waste my time speaking directly to them.
Instead, I found the two domains where their software originated and decided to block them from my network. Amazingly, both my domain, registrar, former payment portal, and even other sites like Ham Radio Outlet became much more usable. The strange focus behavior I found at Namecheap also ceased, which tells me AccessiBe was responsible for that even when their overlay was allegedly supposed to be inactive.
I decided to first share this information on Twitter as keeping it to myself would have been unethical. Later on, after learning of troubling reports of conversations the team had with respected persons with accessibility experience, the illegal activity monitoring they employ, their documented astroturfing of phony WordPress reviews, and the hiring of Michael Hinkson to serve as their token blind man, I decided enough was enough.
At the end of February, I began documenting problems with the AccessiBe software on Twitter and asking their representatives to explain why the overlay was behaving so poorly and why more importantly, the issues had yet been unresolved. The only response I and others received was a handful of pre-constructed tweets from their social media manager.
Finally, having had enough, I compiled my instructions for blocking their overlays into a document titled AccessiBe Gone. I will put the URL in the show notes for those listening on the podcast. If you are listening via some other means, I’m going to try and explain this URL if you wanted to have a look. It is sclower all joined together. It’s S-C-L-O-W-E-R.github. If you’re not familiar with that site, G-I-T-H-U-B.io/accessibegone all joined together, no dashes or anything like that. That is sclower.github.io/accessibegone.
He continues. I published the document on GitHub to ensure it will remain available for those who need it should Namecheap and or AccessiBe decide to retaliate against me.
AccessiBe is one of around a dozen companies trying to sell the same thing to businesses. False peace of mind from the litigious accessibility police. The overlays tactics and general instability give me no practical choice, but to forever ban AccessiBe from my network.
My fiancée and I are free of the plugins incessant nagging and the sites we regularly visit are once more operable from the keyboard. Despite this experience, I am still interested to see where artificial intelligence may be practical to employ across the web, especially since we do have a real problem where the majority of web developers fail to recognize the barriers, their new whizzbang development frameworks can introduce. Sometimes how very little work may be needed to remove them and make a more equitable experience for everyone.
Eventually, I do believe artificial intelligence could have its place in augmenting user experiences. After all, look at what Microsoft and Apple have done with seeing AI, narrator and voiceovers machine learning, user interface interpretation. The key difference between these companies and opportunist tech brothers like AccessiBe are that the former have true talent that understands the needs of the disabled, the wherewithal to implement practical solutions, and also the humility to accept and acknowledge areas that are not yet perfected.
AccessiBe and others like them take advantage of fear and ignorance to sell their products to unsuspecting businesses, showing little or intentional disdain for anyone who finds fault with their software. Any real progress must be made at the browser level. An overlay, assuming it can ever magically fix every accessibility need under the sun would fail unless it were available at all times on a user’s device. That simply is not practical.
Having worked at Ai Squared a few years back, I know this firsthand. We developed an overlay of sorts, which added ZoomText-like features to some high-profile websites. For clarification, I was not on the Zoom bars development team. Despite having actual magnification experts on staff and not overpromising the toolbars capabilities, the product ultimately failed. End users were frustrated by the product’s limitations. A responsibility shared equally between our development resources and the browser security policies.
It clashed with existing access technology. It didn’t work at all on mobile devices. The same challenges are still true with today’s overlays. If the experts couldn’t get it right, what reasonable person would believe vendors like AccessiBe would do any better? As both a user and developer of AT, I find the current trend of silver bullet solutions frustrating and alarming. I appreciate your willingness to jump into the fray and bringing attention to this issue.
If the silent majority allegedly finds these overlays acceptable, then let’s not be silent any longer. Despite my stance, I am looking forward to hearing the alternative points of view that have become a staple of your podcasts. That concludes that email. Thank you so much for writing it, Steve. It must’ve taken some time to do so. We will put some of these issues to Michael Hinkson when he joins us later in the show.
Male Speaker: Be the first to know what’s coming in the next episode of Mosen At Large. Opt into the Mosen media list and receive a brief email on what’s coming so you can get your contribution in ahead of the show. You can stop receiving emails anytime. To join, send a blank email to email@example.com. That’s media-subscribe@M-O-S-E-N.org. Stay in the know with Mosen At Large.
Jonathan: If you’ve been following the discussion about AccessiBe on social media, you’ll be aware that Chancey Fleet is a staunch critic of accessiBe’s technology, their business practices, and the quality and quantity of engagement with end-users. Chancey is a skilled screen reader user, accessibility champion, and her advocacy work includes being president of the tech trainers division of the National Federation of the Blind. I’m just scratching the surface. Chancey, it’s great to have you on the podcast. Thank you so much.
Chancey Fleet: Great to be here with you. I wish it could be under better circumstances.
Jonathan: Broadly speaking, I think that there are two categories of issue. One relates to the technology itself and then the other relates to the conduct of AccessiBe.
I’d like to begin with the tech side. You’ve expressed a view that accessibility overlays are more than ineffectual, they’re actually harmful. Tell me about the harm that accessiBe’s technology is doing and to whom.
Chancey: Overlays in general and AccessiBe, in particular, create harm in a variety of ways. We’ll start with the smallest potential harm, which is that ideally folks bring their own assistive technology to their web browser, into their web experience, and they want their assistive technology to perform in the same way from website to website.
If they’re using magnification tools or contrast management tools or screen reading tools, it’s likely that they want to use those native tools without having to think about doing anything extra. Enabling an accessibility mode and having the website reconfigure itself does a couple of things. It creates a separate and unequal experience for folks. It may change the way that the website represents itself. If you’re communicating across blind and sighted lines, maybe you’re a blind person supporting sighted people or vice versa, it can cause communication to break down if you’re not all seeing the same version of the site.
It can also lead to behavior of your assistive technology that is not the standard behavior on the open web. That just creates a little bit of friction and a little bit of communication difficulties. The second harm that can occur, and this becomes a little bit more serious, is that when an overlay endeavors to solve a problem. As currently constituted overlays don’t communicate back to the developers and guide them to make natively accessible code, whether that’s HTML or something else.
The band-aid gets slapped on and maybe on the user side there’s sometimes a passable experience, but the developer hasn’t learned anything. Throughout their career, they’re going to continue to engage in the inaccessible practices that needed to be remediated both on the website where the overlay’s applied and everywhere else.
That creates a little bit more of a serious cultural harm because even though AccessiBe would have you believe that building accessible websites is more expensive, the truth is that when you embark on developing something new, it really doesn’t take that much more time or expense to make something accessible. Any system that nudges developers away from developing and maintaining accessible coding practices is going to do an ongoing harm. Then the harms get a little bit more serious still.
AccessiBe asserts that they can create 100% automated compliance and they’ll slap some AI image descriptions on some things. When the image is something simple like an advertisement or a dog, that can be fine, but a lot of what happens on the dynamic web is interactive. We might have interactive forums, we might have maps. Maybe I need to know where I can go for a vaccine, maybe I’m looking at a COVID heat map, maybe I’ve just gotten my first Instant Pod and I want to explore the keypad that’s there in a picture. Maybe I’m doing origami. They make fun of me for saying origami, but it is a procedural thing with a complex controlled vocabulary and it’s nothing to make fun of. Anything that’s an infographic or a map or chart or instructional in nature, and anything that has emotional valence, like Bernie and the mittens, any of that is going to get lost to AI image description as it’s currently constituted.
The remediations that AccessiBe says are complete are at best haphazard and mediocre on very standard images. The tool does an all right job recognizing text, but so does my native Chrome image description extension, which is under my control. The tool really doesn’t do a good job parsing anything that’s more complex than might be important to someone’s experience of the web. Maybe I’m looking at a dress online and AccessiBe may tell you that it’s a dress and it might tell you the color, but it’s not going to tell you anything about the detail. It probably won’t recognize a pattern, it won’t recognize the style of the dress. Unless there’s enough alt text there or narrative in a website’s description, I won’t know what it is that I’m looking at. The image descriptions are brutally minimally functional, I think.
Then we get into harms that are, I would say, the most serious. These harms occur when AccessiBe actually makes a website worse. Holly Scott-Gardner’s example, she was shopping for a watch and she turned on the AccessiBe tool, and it turned lots and lots of attributes of each product in a list into a heading level one. As we know, as screen reader users, we move ahead in level one, or head in level two sometimes if we hope to get from parent object to parent object, item to item, entry to entry. If the price of something and the diameter and the description, everything is at that same heading level, it really disables us from using heading levels as an efficient form of navigation.
If we look at other implementations of AccessiBe like on namecheap.com, users are reporting that when the tool’s on, they’re unable to access their dashboards at all, and keyboard focus becomes trapped. I can tell you on another website, I won’t call out my employer today because I’m not here to represent them, but I use a very well-known timesheet service, and there’s another overlay that that service uses called Audio I. Audio I purports to solve all the problems that a blind person might have with the timesheet and I have not been able to use it. It just simply does not work. It may read some content from top to bottom, it purports to reconfigure the page, but it actually doesn’t add sufficient access. This is what we see with AccessiBe as well, what the overlay promises isn’t what the overlay delivers.
Unsuspecting sighted folks procure this technology, they think they’ve done their due diligence on accessibility, they think they’re accessible, and blind folk suffer the consequences. There’s this increased friction when blind folks start to report that a website’s successful and the overlay’s already there, the frontline staff are going to say, “We have this accessibility tool. Don’t worry, it’s already been implemented.” There’s so much more friction to even get the company to listen that something’s not accessible when they’re convinced that they’ve done the right thing by using an overlay.
Jonathan: In your experience, if you don’t want to use the overlay, once it’s been added to a website, is it easy to just completely get rid of it, discard it, decide to use it the way that you did before?
Chancey: It depends. There have been some times when AccessiBe has been relentlessly intrusive and has repeated “press ALT+1 for screen reader mode,” and it’s been overwhelming. In my recent testing last week on the Barilla pasta site, I used AccessiBe with my iPhone and it was easy for me to get into screen reader mode, but when I tried to exit screen reader mode, it would say, “If you exit screen reader mode, the interface would be hidden. Are you sure?” I would click that I’m sure and then it would bring me back to screen reader mode. I was in an endless loop. I think I was able to touch a different area of the screen finally rather than swiping, and get myself out of the Bermuda Triangle of accessibility, but it took me about five tries, and I’m a pretty advanced voiceover user.
Jonathan: Have you ever seen an accessibility overlay making any kind of positive difference? If so, are there certain circumstances where you would agree, even if they are limited, that they can help and they could be a tool in the toolbox?
When AI is being used, it’s really important to have a human in the loop to assess and react to AI results. If AccessiBe were offering a tool that flags accessibility errors, posits a suggestion, and then flags a human in the loop to look at the suggestion, see if it’s accurate and if not, intervene, and if the tool would also flag developers and explain to them how to make the element more accessible natively in the future, I could see an AI guided, more advanced accessibility checker helping to make things better. I think “set it and forget it” automated compliance workflows are never going to create fully equitable access.
Jonathan: How is this different, in your view, from the feature that Apple introduced in IOS 14, where I see a badly behaved app, I can think of one that I completely could not use, and I was really desperate to use this app. Then I turned on the screen recognition features that came to IOS 14 and it performed OCR on the screen. It turned out that once I got past their inaccessible welcome screen, I was fine. Is this a similar thing or are there key differences that people should be aware of?
Chancey: I think it’s very different. Apple created screen recognition to solve the problem locally on the user’s phone, so the user’s not being tracked, their status as a screen reader user’s not being disclosed to any third party. It’s very simple, one rotor gesture to turn it on and off. There’s no hunting down a way to escape from it once you’ve gotten in.
The other thing that’s really different is the way that Apple positions it. Apple is not marketing this to developers as a way to forget about accessibility. In fact, Apple is continuing to invest in the developer community and create frameworks that are easy to use to make apps natively accessible. Apple is honest about the fact that screen recognition is in an early iteration and that performance will vary. Apple doesn’t make broad sweeping claims that extend beyond what screen recognition actually can do. Sometimes screen recognition can recognize images, recognize text, label buttons.
I’ve been able to use some apps, like the Dumpling Grocery app with screen recognition that I couldn’t use without screen recognition. Sometimes screen recognition will actually make an app worse. I have to use just trial and error to see when it’ll work and when it doesn’t. It belongs to a tradition of other tools like OCR and computer vision, that when they run on the user’s side as part of a user’s Swiss army knife of tools dealing with inaccessible content, they’re great, but they’re not put out as a developer side solution that excuses developers from creating inaccessible content.
Jonathan: On the AccessiBe website, they are arguing that expecting businesses, many of which are small operations to pay an accessibility consultant megabucks isn’t working. They say that the current system has demonstrably failed, and they say that with their technology they can get the web quickly accessible and fully accessible by 2025, they think. They say that if you look at the number of inaccessible websites out there, something’s got to change.
Chancey: They cite this number that 1% to 2% of the world’s websites are accessible. I’ve asked them repeatedly to cite that source for me and they haven’t, but some folks think that it might be from the WebAIM study on the top million websites. It’s actually 2% of websites in that top million dataset were free from WCAG warnings and errors, I believe, so that doesn’t mean that 98% of websites are fully inaccessible. That means that there are just errors and omissions on those sites. For example, maybe I’m visiting a site, and the semantic structure of the headings isn’t optimal or data that should have been presented in a table hasn’t been presented in a table. That may or may not lead to an actual usability barrier that’s so serious that I can’t access the web. I would say that most of the websites I need to access on a given day I can fight through, and I’m not excusing the state of accessibility now.
I think claiming that only 2% of the internet is accessible to blind people, I think that that’s a false claim, and it invokes the never-ending task of fixing that 98% where the reality is that if we just iteratively continue to evolve standards and educate developers, particularly we get developers in the educational pipeline and early career to understand best practices around accessibility, and we make it easier for frameworks to natively support and guide developers to do accessible practices, we’re getting a more accessible web. There’s more accessible content on the web right now I can guarantee you than there was 10 or 15 or 20 years ago.
I think documentation around accessibility and the community around accessible development has accelerated the growth of accessibility and I think a lot of the rich web apps that we use these days do have robust accessibility, and for those small and medium businesses that are deploying templates, the solution is not to just play whack a mole and repair things on the client-side, the solution is to hook up with the template providers and develop a more accessible template.
Jonathan: In terms of the way that AccessiBe markets its technology and engages, or doesn’t with the community, following some of the concerns you’ve been expressing I understand that you were one of the people at the meeting referred to in their blog posts and about that they say, “Recently, we participated in some meetings with advocates of the traditional approach. The idea was to bridge the gap between the two approaches, establish that we share the same goal, and are on the same team, and get their perspective on solving the industry’s major challenge which is to make the Internet accessible in the foreseeable future.”
“Unfortunately,” they say, “they were unwilling to discuss this question. Every time we tried to bring the discussion to the main topic, they deflected it by focusing on things like origami, there you go, messaging, and our business model that we are happy to talk about and have addressed, and we’ll address again later on the post, but are far from being the main issue when it comes to solving web accessibility.” They then say that you didn’t have any answers. Could you tell me about that meeting from your perspective?
Chancey: Yes, I’m happy to. It’s a little hard for me to talk about but I’ll do my best. Michael Hinkson initially invited me to a meeting to discuss AccessiBe and he said he wasn’t sure if I’d heard that he joined the company. I replied and I sent him some reading, Adrian Roselli’s and Carl Grove’s articles I believe I sent, and said, “Hey, Mike, I’m concerned about you joining this company. I don’t think we need a meeting. I would urge you to think about what lines you don’t want to cross if you’re going to work with these guys because I don’t think that they’re doing a good thing. I don’t think we need a meeting right now, I’m happy to talk to you about anything else but we don’t need to talk about this.”
He responded and said that my insights would be really valuable and he wanted the meeting more than ever and so I said, “All right,” because I’ve known Michael for almost 20 years, we’re in the federation together and I never want to say no to a federationist who really needs something. Then he sprung on me that one of the founders would be joining and that’s when I reached out to Adrian Roselli and said, “Hey, could you come to this meeting with me to speak from the technical side?” Adrian agreed.
When we got to the meeting, there are actually two founders there. The most memorable was the CEO, Shir. Right off the bat in the first 10 to 15 minutes they were using the word misconceptions telling me that I had misconceptions about their product and saying that they were concerned about my social media posts. I tried to step through some ideas about what AccessiBe can do now and what it can’t do. Among the things that it can’t do is to remediate a table that’s been written into CSS and make that an actual semantic table.
As Holly pointed out, it doesn’t handle headings very well, the image descriptions and form labels are of limited utility and sometimes it breaks access. I mentioned origami because that’s just always something that’s a good example on top of mind for me. I feel like they’ve kind of polluted my relationship with origami because they chose to take that one thing that I said in the span of like five seconds and mentioned it about six times in their blog I guess to minimize me or shame me because it sounds like a silly thing or a niche thing which I guess it is.
I mentioned it in a sentence where I also mentioned infographics, charts, maps, COVID statistics, IKEA instructions, any kind of image where AI is going to fail. I made a couple of recommendations, one of them was that the deceptive marketing has to stop and so this idea of effortless automated compliance has to stop because the tool just doesn’t work in that way. My other main recommendation was that the company do an internal audit, not an audit opened to the public but an internal audit where they could look at the performance of the AccessiBe tool across a range of different content types and elements and make an honest accounting for themselves of where the tool creates valid semantic structure, valid tables, valid image descriptions et cetera. Then bring their marketing and their claims into conformance with what the tool can deliver right now and create a situation where developers can be flagged to pop in and create human in the loop accessibility when the AI can’t be relied upon, and they could use that workflow as a way to educate developers, show developers how the code should look and get humans in the loop to remediate the code, and I think AI could be really powerful at helping with that.
They wanted to focus more on the marketing and the message. When I said to Michael I didn’t want to focus on marketing because marketing was the tip of the iceberg and I wanted to talk about company culture, he said, “I can’t really talk about internal audits or company culture because I’m not there much.” I said, “Michael, what do you mean you’re not there much?” He paused and said, “Well, they’re in Israel.” I said, “We’re all remote now.” I get the idea that he’s been tasked to kind of cajole the critics and massage the message, and I don’t get the idea that he feels empowered or motivated to make any cultural change.
That’s a shame because they are getting a lot of funding, they seem to be smart enough people, I suppose, and they could be working with the community to solve a problem and use a hybrid approach that might actually be of some benefit but they continue to perpetuate these collective harms the tool as it’s currently constituted, breaks access on some sites and on many others. It just puts a bandaid over no wound because the websites are already fairly accessible. It’s just not helpful to us, and it’s not helpful to the businesses that are purchasing something that frankly doesn’t work.
Jonathan: I think what he might say is that you’re attributing motive to some degree there that he genuinely believes that he has seen accessiBe’s overlay making an appreciable difference to the way the blind people can access websites and that’s why he got in touch with them originally to be a reseller, and they brought him into the company because it was enthusiasm and of course, his high profile.
Chancey: I’ve asked the Twitter community and I’ve asked him to point out some websites to me where AccessiBe does a really stellar job. I would love to see it working well on one website that was previously inaccessible and I’ve checked out accessiBe’s own website, it has a lot of WCAG errors, a lot of heading level ones where heading level ones don’t need to be, a lot of links that are broken into a down arrow and then a link and accessiBe’s tool doesn’t remediate those issues, doesn’t really improve things.
I’ve browsed around on Avon and Barilla and a few of their other brand sites looking for a time when AccessiBe seems to make an appreciable difference, I can’t find one and no one else will tell me about one so I’m just not in possession of an AB test where I see that the A for AccessiBe is helping.
Jonathan: AccessiBe have said on their blog that they have been subjected to a disproportionate amount of harassment that they’ve also described what they perceive as hostility directed toward them, and an employee of AccessiBe has talked about the only criticism of excessive be coming from, “Greedy accessibility professionals.” The implication here is that the only people upset about them is those whose gravy train might end because of a cost-effective solution
that’s easy to implement and solve the problem?
Chancey: I’m not on a gravy train. I work for a non-profit and we are not in the accessibility consulting space. I’m a tech educator and my primary motivation is protecting users and businesses from predatory practices or from products that don’t perform. That’s been my alignment all along, all through my career. I’ve never been an accessibility consultant and I don’t have any vested interest in seeing this company or any other company fail. Ideally, I want every company in this space to perform with integrity and effectiveness. They just don’t.
As for hostility, I think, unfortunately, it’s a very familiar feeling to be called hostile. They haven’t called me hostile by name, but they used the h-word in the same article where they said origami six times. If you know me, it’s pretty clear that they’re pointing at me and some other critics. It’s a pretty common tool in the Silicone Valley [unintelligible 01:00:59] culture arsenal to call someone, not always, but often a woman, not always, but often a marginalized person, hostile when they speak truth to power and when they don’t want to engage on the other party’s limited terms.
What I will say is that at no time have I made any personal attacks. I wish Michael well. I think that he needs to find some clarity and bravery about where he finds himself. I think his image and his profile are being used to the advantage of the owners of this company. I do believe that he entered the relationship with the company with some kind of sincere motivation because I’ve known Michael for a long time. I think he’d need to have some fundamental belief in the technology. Maybe they’re really excited about what it would do four years from now.
I don’t know, I’m not in Michael’s head, but he’s not understanding that the company, in 2021, right now, by peddling a product that doesn’t work, is causing harm to our people and to businesses. There’s no hostility in my heart, I just don’t want folks to be subjected to any technology that’s harmful and I don’t want false marketing to happen in the accessibility space. I don’t want people like me and the other critics to have to continue to wear out our voices and use our valuable time defending and protecting folks against this product.
I would much rather we’d be organizing for positive change and building things, and building confidence in the community. This is not a hobby, this is not something that I want to be doing. It’s actually pretty demoralizing, but somebody’s got to stand up and say, “Enough is enough. These overlays don’t work.” Businesses need to be warned.
Jonathan: As an experienced advocate, you will know that, obviously, outlining your case is the first important step. Then, the next step is a clear, kind of call to action, I guess, something that people can do. What do you want the outcome of all of this attention to be?
Chancey: I would love for the major consumer organizations to get together and draft resolutions outlining the harms caused by these overlays. I would love a campaign that educates brands that are either working with these overlay companies, not just AccessiBe, but all of them, about the harms that these overlays do and the untested nature of their claims. I would love for folks to start publishing their own experiences with overlays, maybe in video walkthroughs, so that developers can see what really happens when actual users attempt to use these products.
Of course, I would love it if the companies that currently develop overlays would divest from that strategy and pivot to a strategy that users, AI, and other resources that they may have to suggest and guide remediation and make remediation easier because there does often need to be a human in the loop. I’m not really feeling that any of the overlay companies right now is ready for the transformational change that I think that they would need to become a responsible member of the accessibility ecosystem, but live and hope.
Jonathan: They’ve joined the W3C, so is that a step in that direction?
Chancey: I don’t know. It could be, it could be, but I think that they need to reckon with some of their corporate behavior, pull the deceptive marketing asap, and make a statement to the community that they’ve heard our concerns, that they know where their shortfalls are in performance and that they are committed to retooling their model to align with actual WCAG guidelines, to inform and educate developers and to prioritize preventing the creation of additional barriers for users with disabilities.
Jonathan: Is there any other aspect of this issue that I haven’t covered that you wanted to convey?
Chancey: I think accessiBe’s right about one thing, which is that they are catching a lot more flack right now than the other overlay companies. That’s because Michael, their new Chief Vision Officer, has very assertively pursued meetings with vocal critics. When you get the attention of vocal critics, we engage, we push back, and so the AccessiBe hashtag has been super active. We’re doing everything we can to educate folks about AccessiBe in particular, but we do need to zoom out and realize that UserWay and Audio-I and all those other overlays are also a clear and present threat.
AccessiBe seems to be growing faster and doing more assertive marketing, but we do need to be aware that it’s an entire industry of overlays that we need to confront and that, if we are actually going to prevent collective harm and prevent bad business decisions, we have to be conscious of that in the way that we react and it really does mean all overlay companies, the products that they’re presently offering don’t confer meaningful accessibility. It probably isn’t a great long-term strategy to focus on any one company to the exclusion of the others.
Jonathan: This will inevitably come up at the NFB Convention, won’t it, by way of resolution?
Chancey: I hope it will.
Jonathan: You’ll make it happen, I’m sure. [laughs]
Chancey: Thanks. Jonathan, I’m so fatigued. I almost didn’t take this interview even though it’s really important. I’ve been asked to interview with a few other folks. I can’t tell you how great is the emotional toll of telling a company every day, “Here are some specific situations where your product’s not performing.” I’ve given some really specific examples, including some domains. So have other folks. Their messaging is just relentless and reductive. The strongman arguments that get made and just picking up on origami, the weakest word that I said in an hour of time.
I don’t know if I can be the person to write that resolution, because they’ve got a lot of manpower and a lot of energy. They’re wearing me down, so I’m hoping that other people are going to join the fight, join the struggle and save receipts, document your experiences, create videos. I hope somebody else steps up to write that resolution, because I just– At some point, I have to think about my mental health and think about my day job and just make sure that I’m not in the eye of this hurricane forever.
Male Speaker: On Twitter, follow MosenAtLarge for information about the podcast, the latest tech news and links to things we talk about on the podcast. That’s MosenAtLarge, all one word, on Twitter.
Jonathan: To offer the AccessiBe perspective, I’m joined by Michael Hinkson. Mike is blind himself. Many people know his name as a 9/11 survivor, when he and his guide dog, Roselle, escaped the World Trade Center. In recent years, among other things, he’s been an author, a motivational speaker and he’s spent some years at Aira as well. Good to have you with us, Mike, thanks for coming on. I appreciate that.
Michael Hinkson: Thank you, glad to be here. It’s a privilege and an honor.
Jonathan: Your role at AccessiBe is Chief Vision Officer. What is a Chief Vision Officer and how did you get involved in AccessiBe?
Michael: It’s a title I helped create. Last October, I needed to register a website. We went on to my favorite website, namecheap.com, which I’ve used before. When I got there, this time I heard something that said, “To put your browser in screen reader mode, press ALT+1.” I’ve seen some of those things before. Their plan just didn’t work a lick. They didn’t do anything that made any difference, much less were helpful to me. I went, “Oh, another one of these,” but I pressed ALT+1, because I knew of some things on namecheap that I wasn’t able to easily do.
I pressed ALT+1 and suddenly, the website became a lot more functional and a lot more usable than it ever had been. I was able to read menus, access different parts of the system that I had not been able to access before. I did some looking around. I discovered that the company that did this was called AccessiBe. I hadn’t heard much about them. A little time went by but anyway, eventually, I got back to it in November and said, “I got to go investigate this AccessiBe thing some more.” I did a lot of investigating on the web page and found a lot of claims like, “We make your website accessible,” or “We can do this to your website,” and so on. I also saw this thing that said, “We have a partner program, would you like to be a partner?”
Well, it’s been pandemic lockdown for me since March of last year, I returned from an event in New York where I spoke on March 5th, I came back on March 6th and went into lockdown immediately and I’ve been so ever since. Speaking engagements basically dropped to zero, so I went in and looked and it said if you sell AccessiBe you can get 20% of sales and a variety of things like that and I decided, “Well, nothing ventured, nothing gained, its potential income stream so I’ll try it.”
Well, immediately I was scheduled to speak with someone about the program, and we talked for a while and talked about our histories and talked about what AccessiBe could offer and I thought that it was pretty interesting and so I did eventually completely sign up as a partner and then got connected to Ellie Friedman, who was one of the partner managers, he’s been promoted since he’s a lead in working on a lot of the aspects of helping people engage with AccessiBe well as partners.
Anyway, we talked some, then by January, mysteriously I was getting calls from the founders of the company and after some discussions, they said, “As we really like what we’ve heard about you, we like the questions you’ve been asking, we like what you do and we would like to explore you coming on board to AccessiBe. We’re still in a not really serious income mode, and we need serious income,” so I said, “Well, let’s talk about that.” Because with everyone that I spoke to, I found a staff that was dedicated, committed, open to listening, and wanting to as they describe it, make the internet accessible by 2025.
I say that wanting to point out very clearly they’re not saying, “Make the internet accessible with AccessiBe, they want to make the internet accessible. Eventually, I spoke with Shir Ekerling, the founder of AccessiBe and we agreed on the income prospects and so on and he said, “I want you to be a senior leader in the company,” which was also fine. Again I’ve investigated it a significant amount by this time so I’m not looking at just the product. Of course, the product is will it be usable, can it be better?
More important I was looking at the company because I’ve been involved in several startups including as you mentioned Aira going all the way back to Kurzweil Computer Products when I worked with Ray Kurzweil and The National Federation of the Blind to develop the original Kurzweil Reading Machine. He said, “We have a Chief Marketing Officer and a Chief Operating Officer, what should we make your position?” I asked him what he wanted me to do when he talked about taking the lead and met in helping with messaging.
Taking the lead to interact with consumers, helping us to make this a better product for blind people and for people with other disabilities. I said, “Well then, why don’t we call it Chief Vision Officer?” That’s how the title came about, and a few people I know have made not friendly comments about that on Twitter. Well, that’s fine, they can say what they want. People ought to look a little bit further beyond their nose but anyway, so the bottom line is, that’s how it came to be and that’s how I joined AccessiBe.
Jonathan: Now every company, of course, has its detractors. You can’t be in business without that, but I personally have never seen so many in the blind community so concerned and so frustrated and so angry. Did you know about this depth of feeling before you signed on to AccessiBe?
Michael: I knew some of it. I had heard about a couple of the major detractors, one of whom is not blind who purports to speak on behalf of blind people, and one who is blind.
Jonathan: We’re talking Adrian Roselli in this instance who has written a blog about this?
Michael: Adrian, the first one, and Chancey the second round.
Jonathan: Because we shouldn’t deal in innuendo here so let’s identify those people.
Michael: I didn’t want to mention it unless you did. That’s why. [crosstalk]
Jonathan: I know, that’s fine. Adrian has written a blog post called AccessiBe Will Get You Sued and of course, Chancey has been vocal in a number of spheres as well. You were aware of them before you signed up?
Michael: I read Adrian’s article completely and I found it to be interesting. In fact, I first learned about it because my webmaster, a good friend of mine who lives up in San Francisco, we trade ends of the states. I lived in the Marin county area when he lived down in Southern California and soon after he moved I moved down south. We can’t be in the same place I guess at the same time, maybe it’s a physics law.
David found the article and he said, “You need to be worried about this,” and I looked at it and I said, “Okay, I hear what you’re saying, but I see what I saw when I went to Namecheap.” Also, along the way to becoming a partner, I put AccessiBe on my site and I did that because I wanted to look at my site directly and see what it did. We had just gone through an upgrade, and an updating of my website, we contracted with someone to do it who does not understand much about AccessiBe– No, excuse me, accessibility and he did not do all the things that were really necessary to make my website work efficiently and effectively with a screen reader.
AccessiBe fixed most all of the things that I found that were problems and it fixed some things that I had forgotten were on the site such as images. Now, AccessiBe does not do all images well, AccessiBe won’t necessarily do every image, it won’t do very complex images. Chancey Fleet likes to use the example of, “Well, it can’t describe origami.” My response and accessiBe’s response is, “We never said we could.” AccessiBe is not an end-all solution. The problem that I began seeing as I started working with AccessiBe is that AccessiBe had spent a lot of time and continues to spend as it should lots of time and effort making the product as usable as it can be.
They’ve been focusing on interacting with their partners, the people who sign up to sell the product. They have not spent as much time and I’m going to put it that way, as I would have liked from the outset, working with end-users which is not to say they don’t.
AccessiBe, for example, has had for a few months longer than I’ve been there three different teams, three different focus groups of blind people specifically around the world who are engaged to test the different websites, break AccessiBe, find problems, and report them.
In fact, I’m going to say right now that there was a comment that was made when someone tweeted about finding a lot of functional barriers on accessiBe’s web page and I tweeted back and I said, “I want specifics.” Because what I also started seeing and had happened more after I began to work with AccessiBe, a lot of people are complaining and a lot of people are angry, but we’re not getting specifics.
Jonathan: I’m trying to add some more light rather than heat, and if I can, I’d like to divide it into two categories because I think that’s a very useful summary. The first is the technology itself and then there’s the messaging and engagement. If I can start off with the technology, what venue–
Michael: Okay, let’s stick with the technology, but we’ll come back to that.
Jonathan: All right. Let me ask you the question. What do you think is the value that AccessiBe actually adds? I ask this because I’ve been using a few sites in research for talking to you and talking to others about AccessiBe and what I find is that when you turn accessibility mode on when the AccessiBe plug-in is set up on a site, what I’m seeing is that often there are additional regions that are created which may or may not be helpful. You also see increased use of alt text from images which obviously anybody can license, you could use Picture Smart to do that.
You also see sometimes additional text based on the URL that a link is pointing to. As a blind user yourself, and I guess somebody who represents the company now, what actually is it that you would consider the value add for a blind person, of going to a site with AccessiBe on it?
Michael: If you were a coder, if you were a programmer, and you truly were going to make a website accessible, that may mean in part working to the guidelines of the worldwide Web Content Accessibility Guidelines, WCAG, the W3C World Wide Web Consortium Guidelines and so on, but that does not necessarily make a website accessible. There are a lot of things that are useful in those specifications that are put out, but many coders that I’ve spoken to especially since joining AccessiBe will tell you that doesn’t necessarily fully make the website accessible.
AccessiBe does the very things that you said and it adds value because it produces on your browser and let’s talk about what AccessiBe is and what it isn’t in a sec, but it puts information in your browser that you otherwise wouldn’t have access to. That’s part of the issue that some of the content isn’t necessarily known to you.
Let me give you an example. On my website, if you have AccessiBe turned off, and you tell it to move from heading to heading, it will move to a particular place on the website. I think it says Michael Hinkson, author, speaker, or something like that, but it’s the first heading and it’s some way down the page. If you turn AccessiBe on, and you’re at the top of the page, and you hit Heading, it will introduce an H1, a heading that it calls H1, and it will speak some information that you didn’t know or wouldn’t know was in a heading before.
The reason you wouldn’t know it is because the person who coded the site didn’t code it as a heading. As it was described because I went through this with AccessiBe, the text that AccessiBe is calling a heading is the largest text on a page, and visually, your eye is going to be drawn to that. Whether it’s coded that way or not, you are going to consider that the first major heading on the page, AccessiBe looks at it and says, “This is huge text.” Obviously, there is nothing that indicates that it’s a heading, but visually it would be and so we’re going to make that for the purposes of people browsing and going from place to place on the site, we’re going to call it a heading.
Now, AccessiBe does not change coding on a website. AccessiBe uses this concept that people call overlays. That is to say, AccessiBe is communicating with a web browser and does not change any information on the website itself. Also, AccessiBe does not collect information from you, the user, with your browser. AccessiBe does not exchange information, collect information, or pass on any information to the website owner or anything like that. All AccessiBe does is it enhances the browser experience of people who go to the site and utilize screen reader mode or any of the nine different modes that are available to persons with disabilities.
There is, for example, a mode that one can enter, there’s a way to get to profiles on the page, one of which is dealing with dyslexia, which will for example, when a cursor suddenly encounters a pop up and starts blinking, which could cause seizures, accessiBe’s profile will not allow the blinking to occur, but rather utilize some other method so that people will still be attracted to that area, but not in a way that’s going to cause seizures. Likewise, with a screen reader mode, AccessiBe is going to look at the website and try to pass on information that visually might not be available, such as the heading or non-existent heading as such on my page.
Should that area on my page have been called an H1? Absolutely, the coder screwed up. AccessiBe recognizes what it should be and so it addressed the issue. I don’t have any quarrel with that because it means that I’m getting the same information that other people would get. Now, some people will say, “Well, blind people don’t peruse websites the same way sighted people do.” May very well be true, but it doesn’t mean that we shouldn’t have the same information.
Jonathan: Can I just query the privacy elements around this? Even if I’m a website developer with AccessiBe enabled, and I want to know how many screen reader users visited my site this month, can AccessiBe tell me that even in the aggregate like you had 40 screen reader users come to the site this month?
Michael: No. It’s that simple, no. AccessiBe doesn’t collect any information.
Jonathan: How does AccessiBe know that I’m a screen reader user to offer me the screen reader mode when I visit the site?
Michael: It’s transmitted as alt text and so my screen reader hears it because it knows to look for that text. One of the complaints that people had about AccessiBe and I have it is that if you go to a site like Namecheap, you’ll hear this message it says, “To enter screen reader mode, press ALT+1-
Jonathan: Incessantly. [laughs]
Michael: -and to turn it off, press Alt+0.” Last Thursday, after much discussion, that feature was disabled. If you go to a website and see it today, clear your cache, and it won’t do it anymore because I still kept seeing it a few times and AccessiBe coders said, “Check your cache.” Well, I did clear my cache and that stopped. You will not hear that message incessantly anymore. You’ll hear at once. Now, the problem with that is that some people have said, “Well, but now I don’t necessarily know of that if I don’t hear anything that I’ve gone into screen reader mode.” Well, yes. There are a couple of answers to that. The easy one is just press ALT+1 anyway.
Some people say I don’t want to have to press ALT+1. Okay, I don’t mind but I respect the fact that some people don’t want to press ALT+1. The ideal way to address that is for the screen reader manufacturers to work with AccessiBe to put something into their systems that would allow me to have a setting that would essentially say, “Search for AccessiBe on any website that I go to, and if it’s there turn AccessiBe on.” Now, whether we can get the screen reader manufacturers to do that is of course, going to be a different story, but they should do that.
AccessiBe already has over 100,000 websites that have purchased and use it, including some of those very websites that Adrian Roselli and others say got sued because of accessibility, and I’m not going to speak a lot about all the legal aspects because I don’t want to put any case anywhere in jeopardy.
The bottom line is AccessiBe is a very scalable product. It is a product that a number of people have already found usable. If the end-users that is those of us who go there with screen readers have problems, I’m going to say to you and to everyone listening to this, then email me with specifics.
My email is easy. It’s firstname.lastname@example.org, M-I-C-H-A-E-L-H-I@accessibe, A-C-C-E-S-S-I-B-E.com, I will answer any complaints, issues, or problems that people have. I will get them in conversations with people right now. Shir Ekerling, as I said accessiBe’s founder, wants to be involved in those meetings because he wants to hear it firsthand. He’s not blind, he has no disability, other than being the founder of a company, or an entrepreneurial spirit and that’s not really a disability but sure wants to understand firsthand and really become more aware of the market. His thirst for that knowledge is amazing compared to most people and most companies I’ve ever dealt with.
We’ve been on a variety of meetings with including blind people who’ve had serious problems. I’ve heard a number of stories of people who complain to AccessiBe over the past three years about problems who were very, very intensely angry at AccessiBe, and Shir said, “Let’s work on it and let’s talk about it.” Some of those very people work for AccessiBe today, a number of them are in accessiBe’s focus groups, and we’re looking for more people who have a credible issue, or who can credibly test AccessiBe to address issues that people find.
Again, I’m not going to sit here and say that AccessiBe isn’t introducing a whole new variable of the way, in a sense, we look at websites because it does enhance some of the visual issues like the heading situation that I just described.
I think we need to figure out when people complain, what is really an issue and what is not. I will tell you that what AccessiBe will not do is ignore or push under the rug any complaints that it gets. It may find that the problem isn’t within AccessiBe and we’ll help people recognize that.
Or it may find that the problem is within AccessiBe, and the beauty of finding a problem that is within AccessiBe is that when AccessiBe fixes it, it will address that issue on all websites that use AccessiBe because it’s scalable and everything rolls out through the cloud. A great example of that is www.oreo.com as in the Oreo cookie, which is what the website is. Curtis Chong went to that website and found a graphic that was not being described at all and AccessiBe does try to describe images. For example, on my website, the first time I invoked it and started going down the page, I found an image that I forgot was even there and there was no alt tag with it.
Again, my website updater didn’t do a very good job of coding it and that was him. I’m scrolling down the page and seeing this thing that says, “Man in white dress shirt holding yellow Labrador Retriever.” AccessiBe created that description, which is accurate as far as it goes. It told me exactly what was in the image, sort of like a lawyer, right? They give you all the details and they don’t tell you anything. We did change that alt tag, and now it says Michael Hinkson hugging Roselle, but we left some other ones unchanged.
Jonathan: Mike, you could have got exactly that same information from Picture Smart and JAWS, or even in the newer versions of Chrome. There’s a feature that allows you to automatically describe images using Google’s own licensing. It’s not particularly revolutionary there.
Michael: It isn’t revolutionary, but the fact is it is in AccessiBe. It is there. I’m not saying that AccessiBe is the only way to do that but the very fact is that AccessiBe did do that. We got Picture Smart and JAWS and I’ll get back to Picture Smart because it’s relevant but the very fact that AccessiBe described it is I think something that is pertinent.
It is trying. If you’re a restaurant owner and you are publishing a menu every day and you have an AccessiBe website or AccessiBe on your website, AccessiBe is going to try to interpret that entire menu and describe it to you. Can you do it other ways? Yes. Will they be quite as automatic? Maybe. But the very fact is that AccessiBe put it in the system and I’m not saying AccessiBe is the only alternative and the only solution but the fact is it’s communicating the information to your browser, which certainly is very helpful for a lot of people who may not be quite as knowledgeable about how to use the different aspects of Chrome or JAWS or whatever.
Jonathan: Clearly, this is a work in progress. If I take the Podbean podcast host, for example, which is obviously something I’m happy to geek out on and Podbean has AccessiBe enabled. Its pricing page displays its prices in a way that I would think looks visually like a table but when you look at it with a screen reader, it comes up as a not very helpful list view where all the prices are jumbled up, there’s not much going on and both ways with AccessiBe turned on and off, it renders the same inaccessible ways. What would be ideal is if the Podbean folks actually did the right thing, made a table and fixed that. What concerns people I think is the idea that people feel that they can pay a monthly fee to a company to make the problem go away, but the problem doesn’t go away.
Michael: The question becomes, is it a problem that AccessiBe has or not? In other words, I don’t know whether that looks like a table, but let’s say that it does look like a table. I would also say, for example, if you’re using JAWS, you can configure JAWS so that it will read a row rather than putting it in list view and so sometimes that can be helpful. [crosstalk]
Jonathan: This is definitely not an HTML table. This is displaying as a list. My point in making this example, though, is to say, AccessiBe hasn’t fixed this. It’s clearly an HTML coding issue on the website. Podbean’s probably implemented AccessiBe thinking, oh, good, we’ve taken care of these blind people now and they have not. It’s giving them a false sense of security.
Michael: AccessiBe isn’t going to change the website, it’s doing it and look, if it’s a problem that can be fixed, I think it should be addressed and I’m not defending AccessiBe. What I am saying if Podbean is displaying it as a list of, Podbean isn’t displaying it in a way that AccessiBe can interpret it as, oh, they visually wanted it to look like a table.
Then AccessiBe isn’t going to try to go in and make it something that it’s not, like that heading that I described on my page where it clearly should have been a heading because as it was described to me by several people, it’s the largest text on the page. AccessiBe said it must be that that was supposed to be a heading.
I don’t know how Podbean and HTML coding actually displays it and whether it’s something that AccessiBe can address, but I think it is certainly relevant to bring up and I will bring it up or better yet, if you would email me just so I’ve got it in your own words. I’d like to explore that and see if we can get a better answer.
Jonathan: Yes, but I think you’re not really addressing my concern, which is that Podbean is paying money to AccessiBe thinking that they have now taken care of accessibility. They are under the impression that by paying the what, $49 a month, I believe it is, they can tick the accessibility box and say we are an accessible website now and they are not. That’s selling a product under false pretenses.
Michael: By what definition? The information is there–
Jonathan: No, no. I’ll tell you about the definition. I want to be able to get at the information in a way that tells me very clearly what the price is, how much I will pay each month and how much I will save if I pay yearly. A sighted person can glance at the screen and get that information now. Podbean has paid for your servers under the assumption that you’re taking care of blind people’s access and this is the erroneous assumptions that are surrounding accessiBe’s marketing. They think they’ve done their bit now, so they don’t have to worry about the way that they code their HTML, and isn’t this the danger that they’ve stopped paying attention to the code on their own website now.
Michael: No, because AccessiBe doesn’t say that we’re taking care of all your HTML coding problems that you didn’t do right. AccessiBe is introducing an enhancement that will make a lot of that website a lot more usable. I understand what you’re saying. I’m not sure that Podbean actually comes away with that attitude because it may or may not even really understand what access is all about but now we’re getting to the heart of part of the issue.
Accessibility isn’t necessarily in the coding, it’s in the functionality and the usability. I can’t sit here today and answer why AccessiBe doesn’t do more but I will tell you that I will find out.
If AccessiBe is falling down on its job of making the website as usable as it can and if it is something that can be addressed, the AI can be tweaked. If it is done for Podbean, it will also be done for any other website that has the same characteristics. I don’t know that answer and I hear what you’re saying. There are those that would say, but I can read down the list and get the information and maybe it isn’t as easy as we would like it to be. I’m not saying that’s a good answer, but there it goes saying that, but I think that we can certainly find out what can be done and what the issue is and see what can be done to address it. I’m glad to do that.
You are right. It’s a work in process and it’s going to always be a work in process. Coding is a work in process. I can talk to you about websites that spent $55,000 just to have an audit much less their website fixed by some of the major coders and their coding audit programs say that their websites are 97% to 98% accessible, but accessiBe’s website audit tool ACE, which is @ace.accessiBe.com says that the website is not accessible and in fact, this is a true story that happened last week.
We went with the blind person who brought this to our attention and went to the website. He shared the screen and lo and behold, there is a menu on the site with three items, none of which are accessible, but those three items each open up until about 15 or 20 different things and none of those were accessible. This is a website that was supposedly coded by one of the major experts in the website coding accessibility field. It’s all a work in process. One solution isn’t going to give all the answers, but AccessiBe believes it’s an enhancement and AccessiBe is also willing to say that if you find problems, bring them to us and let us try to fix them.
Jonathan: See, I’m not convinced that that’s the messaging you’re sending on the website. The messaging seems to be that disabled people are predators, who like to sue people and that if you pay this $49 a month to AccessiBe you will make yourself lawsuit proof because you’ll fix the problem.
Then if you go into the terms of service of AccessiBe, it says, and I’m quoting, “The functionality of the AccessiBe systems requires that the licensee website, in which they are embedded, be websites based solely on HTML files and tags.” Basically, no PDFs and things, I totally get that but then it goes on to say, “and that the source code be written according to the standard of the World Wide Web Consortium W3C, without any errors or validation warnings in W3C troubleshooting inspections.”
Now, if on the one hand, you’re saying, we’re going to make your website accessible without all this complex coding, which is way too expensive for businesses, and then on the other, you’re saying, but our terms of service require you to be W3C compliant and write your HTML right. Then if everybody does that, what value are you adding?
Michael: If they do that, but they don’t.
Jonathan: It’s in the terms of service so you’ve agreed to the terms of service when you sign up.
Michael: No, I understand. One of the reasons I was hired was that they began to recognize that a lot of the messaging isn’t what it should be. There are people and I think that it is true on the AccessiBe webpage, people come away with the impression AccessiBe is going to keep you from getting sued. First of all, nothing can keep you from getting sued because lawyers will sue if they want to sue whether it’s a legitimate lawsuit or not is another thing, but people also come away and have said that AccessiBe says that your website will become fully accessible if you use us. Those are not accurate good messages that AccessiBe has apparently projected. Those are being fixed and will be fixed because that is not the impression that AccessiBe wants to give. I think that the WC3 compliant thing is all about writing to standards. It’s fair to say we should write to standards, but we also recognize that most website in this world aren’t written to standards as much as they should be.
Remember, it’s the Web Content Accessibility Guidelines. There’s nothing in and most all of the laws that mandate what the standard is. The problem with standards is standards do not accessibility make. The National Federation of the Blind back in 2008 passed a resolution dealing with web access and accessibility, and what accessibility really means. That resolution says that accessibility is all about functionality and usability.
First and foremost, the guidelines will help with that but the guidelines do not make the website accessible. AccessiBe is using artificial intelligence, not going to be the full solution. AccessiBe is looking at ways to enhance that by having people that website owners who use accessibility can access to address coding issues that AccessiBe can’t, that’s not happened yet. It is something that AccessiBe would like to do because AccessiBe believes that we’re all trying to solve the same problem, and everyone really ought to work together.
The detractors are so opposed to the concept of artificial intelligence being involved, that they won’t and don’t see the value of what AccessiBe can bring. Is it perfect? No, it’s not. Can we make it better? Yes, we can. Will AccessiBe work to make it better? Yes. All the input that it can get, and you gave a specific example today that I’m going to take back to AccessiBe and address and I will get back to you. Then you can pass that information on or we can have another conversation about it.
I’ve heard the comments about the terms of service, as well as the differences between that and what’s on the AccessiBe homepage. I’m not happy with that. We will continue to work to get that addressed because some of it has to be done with lawyers.
Jonathan: Yes, there’s clearly a conflict. One is marketing speak and one is legalese and it’s not a good look.
Michael: Yes, but both of them have to be addressed with lawyers to make sure that we do it the right way. Also, part of it is different languages because AccessiBe is a company based in Israel and very frankly, has not necessarily reached–
Jonathan: Yes, but if you’ve got $40 million floating around, I’m sure they can find somebody to write some proper English, though. Come on.
Michael: That’s right. We’re working. We’re going to get that addressed.
Jonathan: Yes. I do want to come back to the engagement. First, I just want to talk about how some users have said to me that the implementation of AccessiBe has actually made some websites worse. Some blind people say they are constantly being prompted for the accessibility mode. Obviously, that’s now been taken care of. That is a really positive thing. When they do finally relent and enable the screen reader mode though, mainly because they want to get rid of that prompt before you’ve stopped it automatically speaking, users say that in many cases, focus has jumped all over the place.
Some users have said that sometimes accessiBe’s toolbar is actually blocking legitimate parts of the site so they can’t get to it. One listener specifically wrote to me in great detail into the show, saying that he can’t pay his rent anymore, because AccessiBe was installed on the site that he needs to visit to pay his rent. Several blind people have used Namecheap. It’s interesting that they inspired you to join because I’ve seen tweets from people who’ve written to Namecheap and have said, “Look, we’re actually transferring our domains because AccessiBe has made Namecheap less usable than it was.”
We had an example just recently where CSUN has had to issue a public apology for the fact that AccessiBe was implemented on their Assistive Technology Conference site without their consent. It sparked a flurry of complaints and they had to get it removed.
Michael: I’m not sure how that could happen. How would that even happen?
Jonathan: I presume their web host may have done it without their knowledge. I don’t know the details. What I’m getting at is, there is a groundswell of concern about this. Now what’s happening is that you get people on the web who are publishing quite detailed instructions about how to use ad blockers and the settings of your router to actually block accessiBe’s websites and IP address. Now when users experience this degree of degradation and upheaval, there’s a serious problem with the technology isn’t there?
Michael: I think there can be. Let me give you an example of something that we discussed yesterday. I spoke with a person yesterday and Shir Ekerling was on the call, who brought up Namecheap. This person said, “I can’t even log into Namecheap when AccessiBe is on the site.” I believe that person, but I’ve not been able to replicate that. I don’t know whether Shir has figured out what might be going on.
Now, this person used Firefox and I don’t use Firefox. Is that part of the problem? I don’t know. I will be finding out if any progress has been made on that. I don’t know that all of the different things that you’re mentioning are AccessiBe directly related or not. In any way, the bottom line is AccessiBe is affected because people are relating the fact that they’re seeing the site, there’s a problem, and accessiBe’s on the site.
Jonathan: I don’t see how they can’t be related, Mike, because we have two situations that I’m citing here. One, we have a lot of blind people who used to use Namecheap. Now we have a lot of blind people who are saying, “I’m having difficulty.” We had a guy who used to pay his rent on a website, pretty critical task. Now he says AccessiBe is being added and he can’t, so AccessiBe is the only variable that’s changed here.
Michael: What’s the problem? That is, why can’t he pay his rent? What is it that he can’t do on Namecheap? What is he not accessing? In other words, it’s difficult for someone to just say, “I can’t pay my rent anymore,” and we not be able to understand what the problem is. Can he log in?
Jonathan: Don’t you see? The point is he was able to do it before.
Michael: The problem is you can’t solve a problem unless you know what the problem is.
Jonathan: No, no. I’m not disputing this. I want to come at this from another angle. He was able to do it before and now he can’t do it.
Michael: I’m not saying [crosstalk]
Jonathan: What I’m saying to you, Mike, is that this plugin is actively causing harm. I know that the people at AccessiBe take offense to that. Clearly, if you can’t pay your rent anymore, and you used to be able to before AccessiBe came along, you are harming people.
Michael: Again, I am saying to you, without understanding what the problem is that he can’t do to pay his rent, I can’t fix it.
Jonathan: Sure, but if they take AccessiBe out, the problem goes away. Why don’t they just take AccessiBe out? It’s clearly not fit for purpose in that case.
Michael: It certainly isn’t. By the way, they can always hit ALT+0 and turn AccessiBe off.
Jonathan: Yes, but he’s saying that when he does that, the focus jumps all over the place. It doesn’t come out. The problem–
Michael: Not with ALT+0 because that’s what AccessiBe [crosstalk]–
Jonathan: I’m just going to finish this explanation because it’s important. What he then did was he used an ad blocker, such as those you can get from the Google Play, or Firefox, or Edge stores. He disabled the IP address of the servers that AccessiBe used, and suddenly he can pay his rent. Now, that is a pretty compelling argument. You’ve been around enough to know what’s going to happen next. There are going to be blind people who are going to start calling for boycotts of the websites that implement AccessiBe. With stories like that, can you really blame them?
Michael: No, but I want to fix it. I can’t fix it without details. I’m not blaming anyone and I want to fix blame. I also know that a number of the things that we hear from people are, “We’re having a problem. We can’t do this, like on Namecheap.” It took a while to get the person yesterday to tell us that what they were facing was they couldn’t log in on Namecheap. Okay, I appreciate that but I don’t know why. We also got the information that this person used Firefox.
If we can’t replicate the problem, then I don’t know whether it’s an AccessiBe issue or not, but we have to investigate further. Clearly, it is in one way or another been associated with AccessiBe. I’m telling you that AccessiBe wants to address those issues with your person who couldn’t pay their rent, and who had to get an ad blocker involved. I would like someone at AccessiBe to see the actual problem so that we can try to figure out what is going on. Clearly, something’s not interacting from that person’s computer with AccessiBe in the way that AccessiBe wants it.
Jonathan: Now he’s having to go through the hoop of potentially talking to you to get back something he had before. You see the problem. This is where people are saying it’s doing harm because if the company concerned just took AccessiBe away the problem would be fixed.
Michael: You remember, Aira, you worked at Aira. How long was it for us to be able to get Aira to announce again the whole issue of people crossing streets and being able to interact with agents? The problem is no matter how much input we gave, we had a really hard time getting Aira to address that. It finally did but it took over two years.
I’m telling you that I understand what you’re saying, but the problem still exists that coders haven’t done a great job of making products accessible and AccessiBe is trying to do that. AccessiBe may not be doing the job that we want it to do, but we are all part of the same community and can be part of the solution rather than the problem.
All I want to do is to try to really understand by somebody looking at that website, preferably with the end-user and, yes, they have to take the time, but the end-user giving enough information preferably by sharing a screen so that AccessiBe can see the problem and try to fix something that maybe it has to fix.
The bottom line is it isn’t a perfect world today and the reality is that we’re all part of a number of evolving kinds of things. I talked to somebody this morning who said we can’t trust Anthony Fauci here in the United States because when the pandemic started, he said we don’t need to wear masks, but it wasn’t quite what he said. He said that at the beginning the people who really need masks are the people in medical professions and people shouldn’t go after N95 masks, but there wasn’t a major mask mandate initially either. Was Fauci wrong? Hindsight’s a wonderful thing.
Jonathan: You will have seen in your many years involved with technology, people with the very best of intentions who think they’ve come up with the next big thing that’s just going to change the lives of blind people. This is where it seems like the traditional rules of engagement with the blind community as in talking with consumer organizations, tech experts have not been followed since accessiBe’s inception and now they are starting to pay a huge price for that.
Jonathon: When we see these examples of people who they don’t have an axe to grind, they don’t know the individuals personally. All they know is that something they could do before has suddenly become more difficult because of a company that thinks they know what’s best for blind people. Surely in your own career, you have embraced the concept of nothing about us without us.
This technology is flawed and it’s been rolled out and implemented without due diligence and appropriate consultation. That’s why there is such significant backlash and the way, maybe before your time, but the way AccessiBe has responded to that is to play the victim and claim that all these people who genuinely have concern, who have contributed years of their lives to trying to make the web a more accessible place are hostile and somehow have some hidden agenda. That is an appalling way to treat blind people and accessibility experts who’ve done nothing more than try to make the world a better place.
Michael: I don’t think AccessiBe has tried to play the victim. I’ve heard that a number of times, but in my interactions with people at AccessiBe, they acknowledge they have not interacted with end-users, but AccessiBe has for months not responded to articles like the one by Roselli and so on or the videos from Karl Grove, which are grossly misrepresented of reality but the bottom line is I don’t think that it has tried to play the victim.
I think what AccessiBe has done, whether it’s been communicated well or not, or whether it’s totally right or not, I think there’s more right than wrong but that AccessiBe has said there are a number of people who have attacked us without knowing a lot about us. Chancey Fleet said at the very beginning of a meeting that we had that accessiBe’s staff needs to go re-examine itself and needs to decide that it’s doing so much public harm. The staff doesn’t have any reality connection to the world.
She doesn’t know anything about the staff. There have been blanket statements that have come from some of these so-called experts because they believe the only way to gain access is through coding. Even yesterday, the person I mentioned to you who talked to us about Namecheap, when we talked about what’s going to make the world accessible, this person said the only way it’s going to really happen is when coders and when all products are developed with a mandate that accessibility be included. Accessibility again, we finally agreed was based on functionality, not just the standards, but the comment was that it’s not going to be accessible until the time that coders are forced to learn and are mandated by virtue of the development of the products initially to include accessibility and that they can’t be released until the websites or the coding or the product is fully accessible. Great, great goal. No one, including people at AccessiBe will disagree with that. The ideal thing is to put it in the coding.
However, when’s that going to happen? Accessibility is not something that’s going to happen in the near term, but AccessiBe is working to try to enhance the process and bring more functionality to websites sooner than what the coders are going to do. The coders have to some degree, at least, and probably I would say to a large degree, in some cases, plain out now won’t accept that there’s any other solution that can have any meaningful effect other than their coding and their coding hasn’t worked very well.
As I pointed out with the website where somebody paid a bundle of money and it wasn’t accessible, and that was done by a major firm that supposedly specializes in the accessibility business. There is room to address accessibility from a variety of different standpoints, in a variety of different directions. All I am saying is we shouldn’t rule something out that clearly a number of people are finding very helpful, but consumers may not totally be finding useful. We’ve got to fix that.
Jonathan: AccessiBe wants to live in harmony with the accessibility community and I see that they have said that they’re going to join the W3C in the next little while, but you use the term so-called experts. It sounds like by prefacing that with the so-called you are questioning their expertise.
Michael: I’m questioning their ability to truly make websites accessible when they don’t say and they do. I gave you an example of a website that didn’t have clearly delineated menus accessible yet their own audit tool said that the site was accessible. A number of the people who are in this field don’t approach accessibility from a functionality and usability standpoint. They approach it from simply meeting the W3C standards and guidelines and the WCAG guidelines and that does not an accessible website make necessarily.
Jonathan: You were skeptical about my view that AccessiBe has been playing the victim to some degree. I want to read you just a brief passage from the blog post you’ll be very familiar with from them, which says, “Unfortunately, some people are trying very hard to take control of the narrative and have spread a lot of misinformation but not only that, in the last few months, we’ve been going through a non-proportional amount of harassment that is not only unethical, but in some cases, borderline illegal and involves sabotaging our customer’s businesses and trying to harm our digital assets.” Can you give me some specifics about that behavior and who was responsible for it?
Michael: I’ll give you some specifics that I’m aware of. I don’t know whether there will be a lot of cases so I won’t mention names, but there is one person who signed up who is an AccessiBe partner. It happens to be a blind person who found AccessiBe in a different by accident way than I, liked it and signed up to be a partner and he has been approached by one of the major opponents to AccessiBe.
I have not heard whether they had discussions, but the person who approached and said you shouldn’t be using AccessiBe. It’s a bad company and we need to talk about it. It wasn’t just that he was using it, he was a partner. That is not the way to address an issue like that and that does interfere with accessiBe’s business. I will tell you that I was called by a lawyer who said you should run away from this company.
In his basic philosophy is they charge a monthly price and what happens if somebody decides not to use AccessiBe anymore and stops paying it will go away. My response to that is what happens to the company that pays for accessibility from a coder, but doesn’t have a maintenance agreement and something changes on the website? I’ve heard other examples and I mentioned the meeting that we had with Chancey and Adrian Roselli was at it. There were comments made at that meeting that were not acceptable but there are also videos out on the web of people who purport to be demonstrating what AccessiBe
can’t do and AccessiBe as anyone can tell, wasn’t even activated during the videos, there was total misrepresentation. There’s a lot of legal stuff going on and I’m not privy to all of it, but there is a lot of opposition and I don’t mind legitimate opposition where we work to try to address it. I do mind opposition where people make things up as they go along.
Jonathan: If people use ad blockers or other technologies to block AccessiBe technology or they called for a consumer boycott of companies that have integrated AccessiBe, would you consider either or both of those things unethical?
Michael: No, what I would say is if people are having that problem, I want to talk with them. I want to see if we can fix it because AccessiBe is trying to become a better product, no, I wouldn’t consider those other things a little off, they’ve had to do what they need to do to get to websites.
I don’t know what the problem is, but it’s like any other product where I have been involved, whether it be from the Kurzweil reading machine that didn’t read a bunch of stuff at first and we looked at material that people couldn’t read and found reasons why the machine didn’t read it and over time, those issues were addressed as technology and resources permitted.
If people complain and won’t allow us to get the information from them to try to resolve the problem, then I think that is a problem, unethical? No, a problem though, yes.
Jonathan: One of the issues is trust I think at this point because AccessiBe has been around a while now, and you talk about talk, which is really important. One of the things people have been seeking is some open dialogue. There are those who are concerned that they feel that people are being picked off for one-on-one discussions rather than any kind of public forum that is not in any way, screened by AccessiBe, where people can come along and have free and frank discussions with you and share and whoever wants to turn up.
It seems that AccessiBe is not willing to do that, I understand there is a webinar coming up where you can submit questions and those questions will be screened, but there’s no open forum where people can actually have a moment where perhaps some healing could begin.
I think the other thing that probably needs to happen is a bit of a mia culpa, this comes back to the little section I read on your blog. We also see some posts on LinkedIn from one AccessiBe employee, who describes the only objections coming from greedy accessibility consultants, so there’s a lot of rhetoric on both sides.
Don’t you think it’s time that AccessiBe essentially said we are sorry, we have handled our engagement with the blind community and the wider accessibility community incredibly badly, and here’s a 10 point plan about how we intend to do better please can we start again?
Michael: I think we can work on trying to deal with the messaging and address some of those issues, absolutely and I will take that back to AccessiBe. I want to go back to your webinar comment, yes I believe that it’s time to do that. Someone wanted a challenge to AccessiBe to a debate and wanted to have that live, and AccessiBe won’t do it with that individual because they have been very seriously misrepresenting a lot of– [crosstalk]
Jonathan: This is Carl Groves.
Michael: It is Carl Groves and that’s a legal issue and I won’t go there and I don’t know Carl Groves. I’ve seen some of the things that he’s done and I’ve seen some of the things that he said, and he was dead wrong. First of all, the issue with private meetings is only to be able to look at specific examples. I’m perfectly glad to promote and arrange for and get some sort of public forum where people can come and talk about their issue but I want people to come to that with enough of an open mind that if we can resolve those issues– Well, let me put it a different way, why would we have a webinar? We want to have a webinar so that there can be frank discussion about problems, would you agree?
Michael: Okay and if people come to that and want to discuss problems, do we also want to try to move forward from the webinar with either common grounds for trying to solve them, or at least a commitment in this case on accessiBe’s problem, accessibility’s part to figure out how we can address the issues or respond to them
if we can’t do it all within the webinar.
Jonathan: This is why I have divided my questioning broadly into two categories, the technology itself and the messaging and the marketing because this is where the trust comes back. We haven’t even discussed some of the accusations about paying for praise, about besmirching other people’s technology that competes with yours, about WordPress reviews.
There was a lot out there, relating to trust and the burying it under the carpet thing clearly hasn’t worked because if anything, the fact that we’re spending all this time on this podcast talking about it when we focus on issues of concern to the blind community, says it all. It’s reached critical mass, so if it’s got on this podcast, it’s reached critical mass.
Michael: Well, it has, but I don’t think that AccessiBe has tried to bury it under the carpet, AccessiBe hasn’t done what I would like to see it do and what I’m going to work to fix, we are fighting an uphill battle to address it because there is a lot of momentum, but burying it under the carpet and just saying it doesn’t exist, no. Has there been misrepresentations and mischaracterizations of AccessiBe? It does go that way as well. AccessiBe doesnt view- [crosstalk]
Jonathan: But speculation does fill the vacuum, Mike, [crosstalk] only that recently AccessiBe got on Twitter at all.
Michael: I’m not arguing with you about that and again, I joined the company in part, because I was asked to help take us forward and I’m asking people to accept that whether they agree, whether they are happy with AccessiBe, all I’m saying to people is my point is we will demonstrate that AccessiBe does want to try to address issues.
The other part about it is AccessiBe isn’t viewing coders as competitors. If you spend any time interacting with AccessiBe people today, they’re going to tell you that we’re all on the same side and we both offer parts of an overall solution of making accessibility a reality for people of all disabilities and especially in our case, we’re talking about blind people and we want to make that happen but AccessiBe is not saying that coders are competitors. AccessiBe is saying some people have taken a position that there is no other way to make a website accessible other than through raw coding.
Accessibility does not come about because of what AccessiBe is doing simply because of raw coding. AccessiBe also recognizes it is not an end-all solution that’s going to make websites totally accessible and it wants to work with the coding industry, which is why, and I will tell you upfront, I was the one that said to Shir Ekerling joined W3C after Chancey Fleet pushed it at a meeting.
accessiBe’s comment to Chancey was, well, you guys have so demonized us that it’s hard to be accepted and I said to Shir, doesn’t matter, join the W3C, be part of the solution, not a problem.
Jonathan: That sort of language is exactly my point about the victimhood on accessiBe’s part claiming that they’ve been demonized.
Michael: Well but they have, but the point is that AccessiBe moved forward from that and has joined the W3C, I don’t know where all that stands, I believe that they are members or they have done the things to create membership, I do know that there are some meetings that will be occurring and I’ve been told I’ll be a part of it, but I don’t know what they are.
I don’t have the specifics, but AccessiBe has reached out to the W3C and wants to be part of the solution. I agree with you that there’s been some feeling on the part of AccessiBe that they have been demonized. I think it has happened to some degree and I think that this has happened because people aren’t accepting that there is an alternative to doing things by raw coding that can be part of the solution and I am asking you, do you think that artificial intelligence can play a part in making websites more usable?
Jonathan: Well, in this case, I’m here to ask you the questions, but I think it’s a fair question that we should all consider when this episode is over. I do want to, in the time we have remaining ask you about reports that in presentations given by AccessiBe staff, they have distributed eye masks to attendees so that the attendees can experience the “frustrations,” of being blind and using the web.
I take it that you have put a stop to this as a blind guy who was well steeped in the advocacy movement and understands what an ablest deeply offensive behavior that is.
Michael: Yes, I have not seen it, I think I’ve heard it once or twice, but I have not seen it and I wouldn’t like that. I understand the motivation, but I don’t like that, that’s not the way to do it. I would not accept that and I will have that discussion because you’re right, it’s no better than dining in the dark, which is a totally, I would almost say unethical, but certainly, an inappropriate way to get somebody to have an understanding of what it’s like to be blind.
Jonathan: This has been a Frank conversation, and I really appreciate that and I’ve done my best through my research to ask you the questions that I have seen to essentially give AccessiBe. right of reply and while the questions have been tough, in some cases, I do hope you have, at least felt heard, but I do want to finish up by circling back to where we began. Some people have said, if you issued a statement today to say that you made a mistake and now you can see the problems that AccessiBe is causing and that you are going to move on with your life and make a clean break.
Most of the blind community would say, okay. They think that you have made a very grave error of judgment, that your reputation is damaged. I want to give you the chance to comment on this because clearly, you wouldn’t be human if this degree of scrutiny hasn’t taken a toll on you. Do you think you have made a mistake? Do you regret getting mixed up with AccessiBe?
Michael: No. I have become involved with several companies over the years and started my own that weren’t necessarily in the blindness world, but I have become a part of companies because of the expertise I bring from being involved with startups, understanding how startups work, understanding in the case of companies dealing with blindness and blind people, seeing a product, what it might be able to do to enhance our lives as blind people.
I joined companies that have reached out to me and said, we believe that we have a valuable product and some of them have been more dramatic than others and said, we’ve got the greatest thing since sliced bread and many I’ve not joined but if I believe that there is real potential with a product, I believe that if I can make a difference and help it philosophically and logistically become better, I will do that.
I joined AIRA and I joined AIRA with those concepts in mind and with those concepts explicitly stated by the leadership of AIRA. Sometimes they did not follow through on some of the things that they said they would do, like the discussion we had a little while ago that I mentioned about street crossings, where someone at AIRA said, our lawyer said that we can’t have agents talking to people while they’re crossing the street and for more than two years, we fought that battle to get AIRA to change.
Jonathan: On the flip side, I think this is an important point. I want to come back to, you’ll be really aware that even before AIRA launched, Suman Kanuganti, who found that AIRA took the time to know who the key players were in the blind community?
Michael: He did because I took him to the- [crosstalk] I did not have that luxury with AccessiBe.
Jonathan: Right, so he reached out to find people who could do that at a critical time in the organization’s inception and it avoided a disaster because potentially, AIRA could have been a company that portrayed blind people in a negative light that they were helpless without AIRA and because they had the involvement of consumer organizations, particularly in the United States that happened. It’s a bit of a stark contrast from what you’re going through with AccessiBe.
Michael: Right. Unfortunately, that ship has sailed and now we have to work a different way. I’m not disagreeing with you, but I’m saying it doesn’t change what I see as the value of the product and what I see is the potential of the AccessiBe team. Even in the case of AccessiBe compared to AIRA, unfortunately, AIRA hired a number of people who were not blind, who made decisions that should never have been made. AccessiBe is more open to trying to deal with issues now and I can’t ignore what happened in the past, but I can deal with it going forward.
I am only saying to people, if you have any belief in me as an individual, I’m going to do that and you are welcome to reach out to me specifically and I’m requesting you to tell anyone who wants to talk about issues that they are welcome to reach out to me. I will want to understand what their issues are and address them. I am also telling you that the AccessiBe team, right or wrong, didn’t do some things in the past that it should have done. I need to work to address those changes.
I am saying that I joined this company to help it make a difference in a positive way and if I have to turn some things around, I will do that. I believe that we are making progress and that we will affect a major change in making websites more usable by people who happen to be blind. It isn’t going to happen overnight.
It is a different and a new and revolutionary and evolutionary kind of technology. Everything that I have seen, things that I can, and can’t talk about just based on what I’ve seen and being at the company, I believe that the potential is there and that the desire of the accessibility team is to do the right thing. I want consumers to help me define what the right thing is, but I believe AccessiBe is open to addressing that.
Jonathan: Can I just wrap up by asking, is there any point that you would like to make that we haven’t covered so far given that you have been under the grill a bit, I want to make sure that all the key points are made from your side?
Michael: I want to say that I really appreciate the discussion. It has been a good open discussion, and I’m very glad about that. I think the only point that I can continue to make is how I operate. Hey, I’ve got a master’s degree in physics so maybe that governs how I look at things, but for me, I can’t fix something if I cannot understand what it really is and it is going to be a moving target but I am saying that I believe that we can make this a viable solution, other overlays in the past.
DQ systems had an overlay that didn’t go over well and it’s a major coding company nowadays. The fact is that AccessiBe is doing something that hasn’t been done. AccessiBe should have started in a little bit different way by involving more consumers. It is involving consumers. There are several focus groups of people who are hired by AccessiBe to be testers.
Some of those people complained about accessibility and AccessiBe specifically, and what it was, and wasn’t doing. AccessiBe fixed the problems and those people said, all right, we want to be part of the solution. AccessiBe said, we’ll be glad to work with you. If people want to talk to me about becoming a tester for AccessiBe and can demonstrate where they bring value. I’m not going to hire someone just because they say, oh, well, we’ve got this one little problem but I want people who truly can help us analyze.
I will talk with those people and I will tell you that AccessiBe is willing to pay for good solid, ongoing input. I’m not going to sit here and betray blind people. I don’t believe that I’ve sold out. I believe that that it is also time for everyone to take a step back and let’s examine a solution and from my perspective that I believe really does have a way of positively affecting how we all interact with the internet.
Jonathan: That’s accessiBe’s Michael Hinkson, who followed up in an email to say that he talked to Shir Ekerling, the CEO of AccessiBe and its founder, and mentioned the example that I signed it of the website, www.podbean.com, where on the pricing page AccessiBe wasn’t making any difference at all. Shir Ekerling apparently visited the Podbean site and Mike Hinkson writes what he told me was that although Podbean had AccessiBe on its homepage, it was not at all on the pricing page.
Of course, you wouldn’t see AccessiBe do anything to help this page. Several points therefore that, I feel should be included in the interview. One, why AccessiBe is not on the pricing page is for now up to conjecture. My immediate thought after Shir examined the site is that the site utilizes different technologies for different parts of the site and the site owner or their site builders did not choose for AccessiBe to be everywhere. The blog on the site uses WordPress while parts of the site do not.
This is not an AccessiBe issue, but Shir is going to determine who’s sold AccessiBe to Podbean and see if we can learn more about what might be going on. It could be that someone at Podbean doesn’t even know for whatever reason that AccessiBe is not on all pages of the site. Two, as I mentioned to you, there have been many examples of people complaining about what AccessiBe does and does not do on sites.
You cited several examples of specific issues people wrote to you about including the issue with the gentleman who could not pay his rent on a site utilizing AccessiBe. As I said, without more information, it is difficult to really respond. However, you gave a specific example in www.podbean.com. AccessiBe did check this out and identified the issue. No way can access be improve on that page, unless it is first invoked. Three again, I am ready and willing to address specific issues people bring to me, I am working on accessiBe’s messaging and will address issues you raised. I do know that there are some trust issues, some deserved and some not deserved. Like it or not blind people are often testers just as our consumers everywhere with established and new products. Sometimes consumers know they are test subjects and often they do not know and even sometimes consumers are surprised and even possibly angry when they discover that product testing was going on right under their noses.
Even so, AccessiBe does not do the kind of consumer engagement you and I want, and know must occur where the new product, especially where blind people, their lives, and their futures are concerned. As I said, I cannot fix what happened before, but I will address the issues going forward but those issues and perceptions fall on both sides of the discussion.
That’s my appendix. If you will, to the interview we conducted and I will add an appendix of my own, which is that the tone of that email seems to suggest that AccessiBe can be exonerated from the Podbean example that I gave because AccessiBe code was not found on the Podbean pricing page but I think this illustrates the points that a number of opponents have been making.
There can be several reasons for accessiBe’s technology not working. The Podbean example appears to be one where there could be different pages of the site hosted in different ways and that could mean that even though somebody implements the AccessiBe technology on the homepage, it does not filter through to all pages. Even if the technology were working, it may not solve all the accessibility problems. Second, the jury is still out then. I urge you to go to podbean.com and visit the pricing page.
You will see what I mean there, where you can commit to memory the order in which the figures occur and if you’ve got a good memory for those things, you can decipher the information on that page, but it’s not in a nice navigable table and since they have this problem where it’s possible for a piece of code to go on the homepage, but apparently not make its way to all pages in certain circumstances, we don’t know whether it would have remediated the problem, but all other evidence suggested would not have that this technology is not capable of turning data that is best read as HTML tables into HTML tables.
The only answer to this problem is for Podbean to do it right, do the right thing, make your site accessible in a way that’s tried and true. Third, I think that people will have to be the judge about the ethics of blind people testing in real-world environments with something so mission-critical as web accessibility, where if those tests fail, you could be deprived of access to information that you could access before. There may well be blind people who are happy to test for AccessiBe and provide feedback in a sendbox situation.
Many of us are often surveyed for focus groups, various things, and sometimes we’re paid and sometimes we are not and it’s up to us, whether we lend our expertise in any of those situations, but if its own numbers are to be believed, AccessiBe is now on over 100,000 websites. To roll that technology out to that many websites and possibly impede the accessibility of the web while the bugs are ironed out, even if we assume that there will come a moment when this technology does add value could be argued to be highly unethical, profoundly selfish, and playing with the lives of blind people while the money rolls in.
ADVERT: For all things and Mosen At Large check out the website where you can listen to episodes online, subscribe using your favorite podcast app, and contact the show. Just point your browser to podcast.mosen.org, that’s podcast dot M-O-S-E-N.org.
ADVERT: Mosen At Large Podcast.
Jonathan: My next guest needs very little introduction for people interested in assistive technology. I will make a short one. Curtis Chong has been a software developer and administrator in and out of the blindness system and advocate and was for very many years, President of the National Federation of the Blind in Computer Science. His experiences would make a great interview in itself and I’d like to have you back for that, Curtis, but today we’re going to be seeking his perspective on AccessiBe and accessibility overlays in general. Thank you for your time, Curtis. It’s good to talk with you again.
Curtis: I’m very glad to be here, Jonathan. Thank you for having me.
Jonathan: Like many of us, I’m sure you couldn’t avoid the increasing expressions of concern about AccessiBe. I understand that you’ve spent some time now using websites that have implemented it. What’s your take on the technology?
Curtis: My initial take on the technology is neither completely supportive nor energetically opposed. I do believe that the approach that AccessiBe has developed and has clearly implemented across a vast array of websites right now does deserve some careful, thorough, and critical examination so that we, the people who are ultimately affected by this can arrive at a cogent and information-based conclusion as to whether this approach does or doesn’t work.
Now, when I first looked at accessiBe’s promotional material, a lot of it was based on the people who actually pay the money. These are the customers who want to demonstrate that they’re making efforts to make their websites accessible. They don’t have to pay a lot of money to AccessiBe per se, comparatively speaking and the focus of the promotional material was look how easy it was to make our stuff accessible, AccessiBe really helped us out a lot. That’s what I heard from the customer side.
What I did not hear was any commentary from a person, sitting at a keyboard, running a screen reader, accessing websites to determine whether or not for them, the experience was one that was positive or negative. What concerned me more than anything else was this definitely energetic criticism. The first time I sent information out about this to somebody saying, what do you think the words I got were runaway from it as fast as you possibly can.
Being a natural contrarian by nature, I thought, oh, at least AccessiBe deserves a look and go look at some of the websites and see what’s going on. Now there’s a couple of valid points I think that AccessiBe makes about scalability. We’re not talking yet about the pros and cons of what they’re doing. What we’re talking about is if you have people, companies, organizations who are small businesses, who don’t have a lot of money in their budget, they’re using third-party tools or a very simplistic tools to put up a website that doesn’t cost them a lot of money to put up.
If a blind person wants to go to their website, you might experience some problem with accessibility but then if you go to these small companies and ask them to do something as simple as say, for example, putting up an alt tag on an on-label graphic, what will happen is they will say, we don’t know what an alt tag is, or we don’t know HTML. We depend on this other company to do it and so, you’re likely not to get the thing fixed.
You have thousands and thousands of websites going up all the time and we don’t know how many inaccessible sites there really are out there. We don’t know what we’re missing because we don’t go after all of these websites but what we do know is that what’s not happening is that 100% of all the sites that are put up in the world are 100% going to be accessible to the blind.
We know that’s not going to happen given the technology that is in use today, and AccessiBe makes a great point about this. They say that their solution is scalable. They say that it’s relatively easy to implement. What they’re not saying is how much responsibility do we place, or how much emphasis do we place or encourage our customers to place on accessibility so that when the issue comes up, a blind person wants to use a site that’s been developed, implemented the AccessiBe solution.
What is the blind person to do then when suddenly they raise concerns with the company? Suppose an image is described by AI as opposed to a human being and what if the description is somewhat inaccurate, who’s responsible for taking care of that? Who’s responsible when you encounter inconsistent behavior between browsers on your computer? I use three of them, myself interchangeably.
I have a couple of preferences, but I noticed that behavior was different, but AccessiBe, as I said, makes a valid point, which nobody has been able to address, which is scalability. How do we get it to all websites? In an ideal world, what would happen is the tools that you use to develop websites. All of them would develop accessible content. In an ideal world, 100% of all the websites we need to use would work splendidly with our screen reading technology In an ideal world, if a blind user encounters a problem, it would be a place that they can go to call somebody to resolve the issues and not run across a customer service agent who says, go look for the gear-shaped icon, or look for the microphone and then click on that because that’s often what we run up against. The reality is we’re not living in the ideal world, not all websites are accessible. I’m more interested in terms of when I’m looking at what is AccessiBe and others have done, is whether the AI approach is mature enough that it really results in a benefit for us as end-users who, by and large don’t understand what WCAG is? We don’t know the specifics of any of the accessibility guidelines. All we know is that we want to get a little piece of work done, or we want to buy something or we want to get a piece of information. If it doesn’t work, we often have trouble figuring out, why doesn’t it work? How do we describe the problem? Et cetera.
Jonathan: There’s the rub, isn’t it? If we look at the technology as it stands at the moment, I’ve had listeners email me, for example, to say that after AccessiBe was implemented on a website. They’ve actually been locked out of being able to pay the rent, do things that they could do before AccessiBe was implemented on that site. I think we all agree on the problem, which is it is difficult to scale. Is this the answer to the problem, though? When I’ve been having a look at some sites where AccessiBe is implemented, I must say the changes they are making appear to be quite rudimentary.
Sometimes they will reformat the web page and introduce some additional ARIA regions. Not necessarily for the better, I must say. They are licensing some technology, where you do see some additional photo captioning, but you could do that with JAWS and Picture Smart. They’re also giving a little bit of extra information about the URL that is pointed to, in situations where that might be ambiguous in the ALT text. Again, you could do that with a screen reader these days as well. They’ve articulated a serious problem? Do they actually have a solution that is adding value, though?
Curtis: Well, that’s a good question. I will say that, if for example, the problem you’re experiencing with a website is that the menus don’t expand as you expect them to when you hear JAWS say menu collapsed because you haven’t activated the AccessiBe interface. You push Enter or the Spacebar and you’re trying your darndest to get it expanded. If by activating AccessiBe, you actually can expand the menu with a keyboard command. That’s one indicator. Mind you, I’m not here trying to sell the virtues of AccessiBe.
I have run across the same issues that you’re talking about. It worries me that somebody can get up and say with validity, that a task that they used to be able to do is no longer possible with AccessiBe running. The one good thing about, I suppose the way in which the implementation has been set up is that you can deactivate it with a keystroke. We’re focusing on the Windows platform. I did ask them how they were doing in terms of if you were using Safari, let’s say on an iPhone, for example. I did play with a couple of websites using Safari on the iPhone, and the behavior is definitely different.
There is no keyboard command to activate or deactivate the interface. There’s a button at the top, if you can find it. I have identified a few problems, which I duly reported to AccessiBe knowing that a website was an AccessiBe-supported site. You can’t tell just by looking at the web page with your screen reader, whether a site is AccessiBe or not. You can only guess based on the fact that you know it says you have to hit Alt+1 or Alt+0 or whatever. I’m a T-Mobile customer here in the United States. The T-Mobile website has you press ALT+0 to bring up what they refer to as their accessibility menu.
I initially thought that was AccessiBe and found out later on that, that was not. My take on the way AccessiBe is doing this is number one, they’ve got a large number of customers now. That goes without saying, they’re paying them good money and they’re a growing corporation. They have done this by persuading these customers that their work will help their customers to be accessible or at the very least, to avoid lawsuits. I think any idealistic claim in that direction is misguided. I do not believe honestly that AccessiBe will be able to fix every accessibility issue that the customer runs up against. without some extra work or responsibility being taken by their customers.
I’m just simply not going to go that far. I do believe that their acquaintance with the way blind people use the internet, the way in which we browse the internet, how our screen readers specifically behave with different browsers is something which I regret to say has only come to their attention lately. I’m not saying this as a way to say Oh, AccessiBe is a horrible thing but the way they’ve dug into the hole, the end-user benefits from their work, in my view has been insufficient. They need to do more. They’re starting this, I think but they need to do more to understand okay, we’ve put this thing in how does it look? How does it sound?
How does it feel with respect to your computer, your situation for the specific websites that were supporting? Which there are a lot of them. I’m not a person who wants to test thousands of websites. I don’t have that kind of time. I got lots of other things I wish to do. I hope they have other people who are doing it, who were working for them because if they don’t, they’re making a huge mistake. That concern I have has nothing to do with technology. It has everything to do with implementation. How do you help your customers? I’m saying this from the AccessiBe perspective to understand that if a customer says, my responsibility with regard to AccessiBe is done.
I’ve paid for AccessiBe, they’re supporting my website and I don’t have to do anything else about it. Any problems you have with our stuff, now go call them.
Jonathan Yes. See, this is the thing because people are paying for this. They think that’s made the problem go away. Particularly with the marketing that’s on their website, it’s almost like, inoculate yourself against these disabled people who want to sue you. In one of the testimonials, they talk of frivolous lawsuits. I think there are two strands that you have identified that are very important. First, there’s some doubt about whether the technology really is adding value? Whether in some cases, it’s making life worse for disabled people?
Then on the other hand, there’s something you’ll be very familiar with. This is the lack of engagement with users, particularly at the conceptual stage. I’ve no doubt that you will have had many people come to you with the next big thing. You’ll have done a lot of educating. As a product manager, I’ve been there too. People have done a very minimal study and they think they’ve got this amazing thing that’s going to change the world for blind people. Then they actually take the time to talk to a real blind person, what lessons do you think that AccessiBe can learn about engagement that the rest of the community might benefit from?
Because honestly, I think one of the problems that they face is the real poor quality of their engagement. They seem to have upset a lot of people by not taking people’s concerns seriously.
Curtis: Some people tell me that not taking concern seriously has been a growing, growing problem. From my perspective, I only heard about this at the end of January. I came to this lately and not engaging the end-user during product development is a concern that I have across a lot of things that happen to us in the world. How often do we hear about wonderful products that are developed on our behalf, but they’ve never talked with or continue to work with real living, knowledgable blind people during the design phase? Which is when specific changes can be made without ruining the end result?
I think I get the feeling. This is not substantiated by too much research on my part but this is the first time I’ve heard of AccessiBe. If they’ve been around for three or four years, it seems to me that they should have been talking to us. They should have been making a lot of noise about engaging with the blind community, which they haven’t adequately. Because of that, now they’re going to hear lots of people complaining about their stuff. Some of those complaints, many of those complaints, let me say, will be justified. They will be something that AccessiBe will need to pay attention to.
If what they’re doing now is writing on their laurels and saying, look, we got all these customers. We’re making all this money. We’re protecting them against lawsuits, that we’ll even go to court to defend them. Without talking to the blind community, to us, in general about what needs to work? What doesn’t need to work? How is it working? How is it not working? That’s where they’re going to incur more resentment, which would have been resolved earlier if they had done this earlier. I’ve said on multiple occasions to different people, they are doing something which they should have done two or three years ago. That engagement with the end user is the ultimate test in my mind.
Jonathan: You made a comment I want to come back to about disabling AccessiBe by pressing Alt+0 on websites where it’s enabled. Some of my listeners who’ve written in already on the subject have said we’ve tried that. Yet, there’s still elements of the accessibility toolbar that hang around on the screen.
Curtis: Yes, I found that also.
Jonathan: It’s obscuring some content. That’s where people are getting into trouble. Also at least historically, there has been this repeating message that you keep being interrupted with I guess, a live region of some kind that keeps telling you to enable accessibility mode by pressing Alt+1, which stops your engagement with the site.
Curtis: What’s worse, Jonathan is you don’t know definitely whether AccessiBe is engaged or not engaged when you return to a website? You’re in doubt. There’s no clear unambiguous indicator that says, hey, we’re working or we’re not working. Sometimes you see a button that says, screen reader mode is active. Other times, you don’t see it. I saw in YouTube for example, where they said well, we have links at the top of the page that say, go to header, go to footer, go to the main part of the screen. I look for those, but I found that those vestiges, you’re absolutely right, stayed there after I turned off the AccessiBe interface. There are issues. These are technical issues, by the way. I’m going to say it straight up. They should have caught these a long time ago. A very long time ago.
Jonathan: Yes, it’s a lack of engagement that’s caused that. But where I was going with this is, you have people now who say the only thing that they feel they can do to get on with their web surfing life, and we are talking about people who use certain websites for their livelihoods. This is not just a consumer recreational type-thing we’re talking about.
Some people have actually published instructions on websites now that explain how to use ad blockers and software built into your router to disable the IP addresses of the accessiBe’s server so that you don’t have to engage with it. Now, when people go to all that trouble, and it does take some trouble to do that stuff, you’ve got a real PR problem there if you’re an AccessiBe person because it’s being treated now by some members of the blind community as malware.
Curtis: Now, here, Jonathan, I think you’re exactly right. You, of all people, have done the most, in my mind, to help me understand what is the problem that people are encountering on a day-to-day basis with AccessiBe. In other words, I have heard many generic rhetorical commentary about how bad the approach is. That’s not what I’m looking for. I’m looking for exactly what you provided.
Jonathan: That’s why we’re trying to do the show, by the way, is to add more light than heat on the subject.
Curtis: Right. The one way that I found to disable the AccessiBe interface for sites that I have previously visited is, of course, you have to essentially delete everything, your cookies. All that kind of stuff. That seems to be the only way to remove it. Now, I regard these problems as serious bug problems that require attention. In one sense, I think there’s an interesting idea. Maybe the technology isn’t there yet.
I don’t know, but you go to a website that nobody else visits, whatever, and you find a problem with that website. Let’s say it’s not an AccessiBe customer. You have no idea who you’re supposed to go talk to when you encounter a problem with that site. Now, let’s assume that AccessiBe was working perfectly, that we didn’t have an issue, that the AI was great. This is down the road, right? Because I’m not saying it’s perfect now.
If I run into trouble with a website knowing that it’s supported by AccessiBe, I could have, let’s say potentially, a place to go to who could, potentially, be on my side who understood accessibility and who could say, “Yes, we think that’s a big problem. We’ll go through and we will fix it.” If that’s what we get out of this, and if it really produces something that really works for us, I can get behind it.
But I’m not ready to do that now, because I think that there’s enough imperfection here that requires attention that needs to be addressed. May I also say that that concern is equally true in my mind for websites and entities who are claiming to be accessible using the traditional methods for making their stuff accessible. How often have we dealt with companies where they say, “Yes, we meet the WCG guidelines.” We’ve got version one.
It’s working great. The blind people try it. They say, “Yes, it works great.” Then nine months from now, when the team has left the company, a new thing rolls out, and all of a sudden everything breaks. This happens to us way more often than it needs to happen. We are spreading our energies across multiple companies and organizations today because for each accessibility issue we run across, we know we have to go directly to that company to try to fix it, and often our life has changed.
Let’s say, for example, we have a great deal of money in a bank, and now the bankrolls out something new online, and we’ve done everything online and it’s been working tremendously. Next year, they roll out a new version. All of a sudden I can’t do my business anymore. I can’t pay my bills. I can’t do whatever. What am I supposed to do? The worst-case example is I have to leave the bank, take all my money out, grumbled to the manager who won’t remember why I did it, won’t communicate that to the upper office and I go to the other bank that will get me what I want.
That’s what happens in today’s world. Potentially, the AI approach could work if it were done correctly, but so much of the success or failure of this depends on factors that are not technological. They are human interaction factors. They are things like, do you engage with the blind community early? Do you have people on staff who really understand how screen readers work and interact with browsers?
Do you really accept that you’re not just dragging in some average blind person who doesn’t know very much about technology, but who is somebody you’ve made contact with and you think, “Gosh, I’d like to hire that person, but they know nothing about how to navigate the websites with their screen reader, et cetera” to advise your developers, et cetera, et cetera?
If you do that kind of stuff and you just do surface involvement, which happens way too often, the approach, the method, or the system is going to be a disaster for those of us who don’t work for the company, but who ultimately either benefit or lose based on the effort that’s been done. I’m at the stage where we need to hear from those people who have trouble with AccessiBe.
They need to be paying attention worldwide about this because they are an international company and they need to really engage with the community, period. The community, on the other hand, those people who can do it, need to gather up specific, problematic things that aren’t working and get that information into the hands of AccessiBe. Because right now they haven’t done their work right over the years. Now they need to start hearing from people about specifically what’s wrong with their system.
Jonathan: There are those who say, though, that this technology, and I’m talking about these accessibility overlays in general, not just AccessiBe, are intrinsically harmful. It sounds like you’re not saying that. It sounds like you’re saying there is capacity to work with these people.
Curtis: I felt the same way about AIRA, to be truthful. When I first started to feel about AIRA, I feel the same way about all new technologies. I look at it with a great deal of skepticism at the beginning, and then after it has proven itself as something which actually has more benefit than drawback. I don’t know whether the approach has more benefit than drawback or more drawback than benefit yet.
I think it needs to be exposed more to the people who were ultimately going to supposedly benefit from this so that we can gain. So far, we’re seeing signs in both directions. Yes, some benefits, some costs. I think we need to do that without a lot of what I call rhetorical criticism either way. When I first looked at AIRA, that’s my practical example of how I got pushed into something despite my reservations to the contrary.
I was very worried that newly blind people or people who had a very low self-confidence would find it something that they would become overly dependent on. I worried about that but after a while, I stopped. Because I recognized that the key was how it was implemented, not the technology itself. I think that’s true with AccessiBe and similar approaches. Do they hurt us more based on specific examples or do they not?
Do they have a mechanism to hear from people? I think of going to the web in a lot of situations as I visit a new neighborhood at least twice, three times a day, maybe more often. How do I know if this is an AccessiBe site or some other? What do I do then if I have concerns about the interface not working properly? Heck, I can’t even figure out who to go to when I have a real accessibility issue on a website that has none of this stuff implemented in there. Now our problem has become worse.
Jonathan: We can all relate to having an issue with a particular application and you contact your screen reader vendor. They say, now you need to talk to the vendor of the actual software that you’re using and they sent you back. No one takes responsibility. I want to come back to the AIRA example and full disclosure and disclaimer, and all that good stuff. Of course, I was directing communications for AIRA at one point, but I think it’s an interesting parallel to draw because AIRA engaged in a very meaningful way with NFB and other organizations even before they launched a product.
They took the time to say who are the key people that can give us advice and I think you’re absolutely correct. AIRA had the real risk of going way off the rails and portraying blind people as dependents who can’t survive without their technology. It would have been in their commercial interests to do just that but because they got not just advice, but also investment from NFB and others, they didn’t do that.
They actually engaged with a number of blind people, including in senior leadership positions who would stop them from doing that. That’s quite a contrast from what we’re seeing with AccessiBe.
Curtis: Correct. They’re only coming at this lately now, and they’re going to pay the price for it. They should have done a better job of researching the community. I can understand why they did it and the reason was they weren’t trying to hurt blind people. They didn’t come out there and say, “Yes, let’s just not talk to them because they don’t know what they’re talking about.” They went after the revenue, the money.
“Who will pay us to do this great work that we want to do and how do we do this the best?” They focused on that. Great. Terrific. I’m all about companies making money, but I don’t want them to make money at my expense. They can remedy this, but they’re going to have to endure a lot of criticism because they failed to come to talk to people early enough in the process. If I were them, I just grow up, take the criticism, fix the stuff, and justify the support. If that’s what they deserve, show us that what you do really does work but let’s try to do this in a way that doesn’t result in flame wars all over the place where people are yelling at each other, because some people think, let’s say access or the approach of AI is the best thing ever and other people think it’s the worst thing ever.
ADVERT: What’s on your mind? Send an email with a recording of your voice, or just write it down. Jonathan@mushroomfm.com. That’s J-O-N-A-T-H-A-N@mushroomfm.com or phone our listener line. The number in the United States is 864-60 Mosen. That’s 864-606-6736.
ADVERT: Mosen At Large Podcast.
Jonathan: Roberto Perez [unintelligible 02:51:04], you are taking on a big topic this time. I think the problem is not so much the overlay, but the fact that products such as AccessiBe are more than often referred to as a solution for web accessibility. When they are used as a short-term solution and intermediary step, as part of a larger process to make a product really accessible, they can have a place. Unfortunately, due to lack of awareness and irresponsible marketing, they are presented as the ultimate solution.
The root cause of inaccessibility is not the technology, but the lack of awareness for culture that leads to the creation of technology that is not accessible. Even if products such as AccessiBe can resolve some problems, such as keyboard interactions for accessible controls, they create an even bigger problem in the long road by taking the responsibility for web accessibility out of the hands of companies to put them in the hands of the creators of the overlay.
As customers with disabilities, by accepting this solution, we give away the possibility of a more inclusive society in exchange for a few technical fixes that more than often are not even enough. I also think that business owners can save more money by investing in educating the product team, which in the long road can result in a better product for everyone, not only disabled customers.
Turning to a comment now from RJ Mayer, who says, any time you use or have to click a special pattern or command in order to have equal access to a website, it is discrimination. A user should not have to out themselves as being disabled to have equal access to a website. It is discrimination. No amount of marketing can take away that fact. If an overlay did its job, then it would work seamlessly with the web code and not force the user to be segregated. This one does not.
False promises made with slick marketing do not create an accessible world. Steve Cutway writes, we employed an accessibility overlay at Queens in 2005 whose name, I can’t remember, but I think the product had transcoder in its name. The company was from New York City and they visited us frequently. Web design and screen reader improvement since have made the need for such overlays unnecessary in my view but I understand why they are attractive to business.
The number of websites that are truly accessible is in fact, very small. Now that the web is business-centric, the wild and free web days you pined for in the February 6th, 2021 podcast have gone, I’m afraid. I went to the AccessiBe website and tried it with, and without screen reader mode. I know that I’m probably not the typical web user having used it since its early days in the 1990s with Netscape, but I didn’t find screen reader mode enhanced my access in any way, particularly as it appeared to get turned off as soon as I clicked the watch demo button.
I do see the attractiveness for businesses and others who were using third-party developers who probably have never confronted accessibility. I don’t know the first thing about it and may neither have the time nor the inclination to learn. That’s a reality. From Rebecca Skipper, I welcome any efforts to make websites more accessible, but we need to look beyond consumers and the visually impaired and consider those who were working and have other disabilities. I have seen overlays on a variety of websites, including the social security administration.
Using overlays is just as bad as offering a text-only version of a website. I thought the purpose of the WCAG guidelines to encourage universal design. By focusing on the screen reader users, we are leaving out people with dyslexia or other disabilities, including low-vision magnification users. Developers should know about accessibility and know-how to create accessible apps. No one should be dependent on one tool.
That being said, if an overlay is the only way to make a website accessible, then it is better than losing a job because you could not use an online tool. Thanks, Rebecca. AccessiBe does offer a range of options, which you will see depending on the technology you are using. It isn’t just a screen reader tool, as you will have heard in the interview with Mike, which of course came after you sent that email.
Kylee: Hi, Jonathan, Kylee Maloney here, protect your business from lawsuits, open your website to the world. It sounds overly dramatic but to me that’s not what accessibility is supposed to be about, while I understand that it has become a compliance issue perhaps in the USA with the section 508. When I have ever educated about accessibility, I have talked about design paradigm and philosophy, and logical structure, and the many reasons why accessibility is a good thing besides doing the right thing, such as the increase in SEO and that kind of thing.
This concerns me somewhat because if you try to make someone do something out of fear of the consequences of not doing it, and instead of giving them reasons why it’s a good thing, a good thing for their business, as well as a good thing for the people who are going to benefit from it. Surely that’s to better in the long run. I am very dubious about this kind of thing, especially at $49 a month, that’s not going to be a small cost for a small business.
Accessibility to me should never be a tick, the box exercise. The other reason is it’s a changing field, new technologies, new techniques. If you try to capture it in a tick the box exercise like this, you’re never going to get accessibility, just compliance with a few static standards. As we know the standards aren’t static, they can’t be, but the philosophy is the paradigm is so let’s work on that and stop trying to scare people into doing something which we need to convince them to do for their own good and for the good of all.
Jonathan: Thanks, Kylee, and I’m sure that this discussion will continue in subsequent episodes of the show.
ADVERT: To contribute to Mosen at Large. You can email Jonathan that’s, J-O-N-A-T-H-A-N@mushroomfm.com by writing something down or attaching an audio file, or you can call our listener line. It’s a US number 864-60Mosen. That’s 864-606-6736.
ADVERT: Mosen At Large Podcast.
[02:58:22] [END OF AUDIO]