The Federal Trade Commission will require software provider accessiBe to pay $1 million to settle allegations that it misrepresented the ability of its AI-powered web accessibility tool to make any website compliant with the Web Content Accessibility Guidelines (WCAG) for people with disabilities.
The FTC press release includes this subhead: Company also failed to disclose material connections to online reviewers
Essentially the FTC found that accessiBe had been misrepresenting itself to end users, customers, and the community. This finding may not be a surprise for many disabled people who have encountered accessiBe in the wild, as evidenced by tools and guides which have popped up to block the accessiBe overlay.
I Told You So?
Readers of this site are probably familiar with plenty of these false claims and misrepresentations because I have been tracking them since 2020 in the painfully long and stove-piped post #accessiBe Will Get You Sued.
This paragraph from the FTC release is a good summation of that post’s 23,000 words:
[D]espite the company’s claims, accessWidget did not make all user websites WCAG-compliant and these claims were therefore false, misleading, or unsubstantiated, in violation of the FTC Act. In addition, the complaint alleges that accessiBe deceptively formatted third-party articles and reviews to appear as if they were independent opinions by impartial authors and failed to disclose the company’s material connections to the supposedly objective reviewers.
Vindication?
You might think this feels like vindication, but it does not.
Sure, it’s nice to be validated by the FTC after 5 years of shouting to anyone who will listen that accessiBe is a net harm for the disability community. But the settlement is a pittance.
Reports put accessiBe’s 2024 revenue at $51.3 million. The settlement has a fine of $1 million. Subtracted from its 2024 revenue, that leaves accessiBe with $50.3 million dollars for the year. Dollars earned by lying to customers, misrepresenting itself to the community, and arguably harming disabled users.
In other words, the financial incentive to continue to be a bad actor is still there. Companies like AudioEye (projected $35.2 million in revenue), UserWay (LevelAccess), and each one listed on the Overlay Fact Sheet simply need to look at their revenues and calculate if $1 million is the cost of doing business.
Only Proposed?
This settlement is proposed. It’s not final. It will get published in the Federal Register and then will be open for 30 days for public comment. Then it goes for final approval. I am not a lawyer.
Regardless of whether it goes up on Monday or later, 30 days brings the close date well past 20 January 2025 (inauguration day). This draft FTC statement on the public comment period is the only reference I have, but will update this post when I know more. I am still not a lawyer.
The good news for those impacted by accessiBe’s history of false claims is that they can leave comments with their own experiences, opinions, demonstrable facts, and / or thoughts on the size of the fine. This will all become part of the public record for this case and available online. I continue to not be a lawyer.
But Legal Force?
This presumes the FTC will have the teeth or the will in the coming administration. That includes the will to enforce the part of the settlement where accessiBe is barred from making the claims or turfing the astro that got it into trouble.
Some good news on that front is that the two Republican FTC commissioners (Andrew N. Ferguson, who clerked for Thomas, and Melissa Holyoak) made it a point to call out accessiBe’s actions as deceptive in their own separate statement.
They also make it a point to explicitly note their support of this ruling should not be taken as them endorsing the position that the ADA, or the WCAG, require[s] a website operator to ensure that some or all of the third-party domains or subdomains with which it integrates are accessible.
This is neither good nor bad, but worth being aware of if they stay on in the new administration (and then become the most de facto moderate of commissioners as a result of the Overton shift).
Probable Win?
What this settlement does offer, however, is yet more ammo for anyone stuck fighting back against overlay products getting the green light from clients or bosses who only see overlays as risk mitigation against their own misunderstanding.
The FTC considers accessiBe to be deceptive. Anybody considering accessiBe as a vendor should know about this and, with this reasonably visible ruling, hopefully will.
After all, the prior evidence, complaints, and bad press have certainly had no impact on accessiBe’s not-quite-hockey-stick growth over the last few years.
Summary and Exhibits
I encourage you to visit the FTC Cases and Proceedings page for accessiBe Inc. You can download all the exhibits, mostly as tagged PDFs but with some videos, though I am embedding them here as well.
Exhibit I may look familiar to you. It is the video that accessiBe’s former CEO (now president) made to address my post in August 2020 (and on which I commented extensively). I have opted to embed the YouTube version of the video. The captions are the same.
Consent Agreement (257.29 KB)
Complaint (1.6 MB)
Analysis of proposed consent agreement (114.24 KB)
Joint Statement from Commissioners Ferguson and Holyoak (217.91 KB)
Related Links
There is far more content online about the failures and false claims of accessiBe and overlays in general, so consider this a curated list for those that explicitly cite accessiBe and not just all overlays broadly.
While this settlement works through the FTC review process, accessiBe is asserting that “The FTC’s complaint in no way attacks the capabilities of accessiBe’s technological solutions to ensure that its clients’ websites comply with accessibility regulations in the U.S.” [1]
I think it’s important to note that there is plenty of evidence to support that accessiBe’s overlay product fails to make all sites conform to WCAG (which is referenced in ADA caselaw in the US) [2]. Its own terms of service said as much in 2020:
“[…] embedding the accessiBe Systems in the Licensee Website does not, on its own, fulfill all of the requirements of applicable law in respect of website accessibility (accessiBe does not remediate PDF files or create subtitles for videos, for example).” [3]
As noted in the post about accessiBe’s overlay failures, it also notes that the overlay sometimes introduces WCAG violations. That review is from 2020, the timeframe covered by the complaint. But even in 2024, I was still documenting WCAG violations introduced by accessiBe’s overlay product [4].
Further, disabled users have been vocal about the harms they have experienced using the accessiBe overlay, as covered by Vice [5], NBC [6], Wired [7], and The New York Times [8], among other outlets and on social media. Some disabled users went so far as to create browser extensions to block accessiBe’s overlay when encountered on sites [9].
I ask that the FTC approve the settlement to which accessiBe has agreed.
Because accessiBe is already framing this settlement as unrelated to the efficacy of its product, I also ask the FTC to mandate that:
A. accessiBe acknowledge the feedback from the disability community about the harms they experienced from its overlay product; and
B. accessiBe acknowledge in its marketing that its overlay product has not and does not make sites conform to WCAG nor any laws, and using it may introduce risks under those same laws.
## References:
1. “FTC Fines Israel’s AI Web Accessibility Company AccessiBe $1Million Over Deceptive Claims”, Jewish Business News, 6 January 2025, https://jewishbusinessnews.com/2025/01/06/ftc-fines-israels-ai-web-accessibility-company-accessibe-1million-over-deceptive-claims/
2. “Bolt-on Accessibility – 5 gears in reverse”, Steve Faulkner, TPGi, 13 May 2020, https://www.tpgi.com/bolt-on-accessibility-5-gears-in-reverse/
3. accessiBe Terms of Service, captured 10 December 2020, https://web.archive.org/web/20201210230935/https://accessibe.com/terms-of-service
4. “#accessiBe Will Get You Sued”, subhead “accessiBe Still Introduces WCAG Violations”, by me (Adrian Roselli), updated 18 January 2024, https://adrianroselli.com/2020/06/accessibe-will-get-you-sued.html#2024StillBroken
5. “People With Disabilities Say This AI Tool Is Making the Web Worse for Them”, Vice, 17 March 2021, https://www.vice.com/en/article/people-with-disabilities-say-this-ai-tool-is-making-the-web-worse-for-them/
6. “Blind people, advocates slam company claiming to make websites ADA compliant”, NBC, 9 May 2021, https://www.nbcnews.com/tech/innovation/blind-people-advocates-slam-company-claiming-make-websites-ada-compliant-n1266720
7. “This Company Tapped AI for Its Website—and Landed in Court”, Wired, 11 November 2021, https://www.wired.com/story/company-tapped-ai-website-landed-court/
8. “For Blind Internet Users, the Fix Can Be Worse Than the Flaws”, The New York Times, 13 July 2022, https://www.nytimes.com/2022/07/13/technology/ai-web-accessibility.html
9. https://www.accessibyebye.org/
FTC Decision and Blog Post, 24 April 2025
The FTC voted unanimously to approve the final consent order.
The final order bars accessiBe from representing that its automated products can make any website WCAG-compliant or can ensure continued compliance with WCAG over time, unless it has the evidence to support such claims.
It also prohibits accessiBe from misrepresenting material facts about its products and services to consumers and from misrepresenting that:
statements made in reviews, blog posts, or articles about its automated products are independent opinions by impartial users;
an endorser is an independent or ordinary user of the automated product; or
an endorser is an independent organization providing objective information.
It’s advertising law 101: before you claim your product can perform a certain task, you need evidence it will work as promised. If your product doesn’t live up to the hype, you’ll have to deal with the consequences—which could include a potential FTC lawsuit. That’s the lesson from a recently announced settlement between the FTC and accessiBe, a company that advertised its online plug-in as a fully automated tool to make any website compliant with the Web Content Accessibility Guidelines (WCAG) […] [T]he plug-in didn’t work as promised. And the FTC says accessiBe doubled down on its problems by disguising paid reviews and articles the company drafted or edited as impartial third-party reviews.
Feels like a damning statement. One that’s hard to dismiss as a current administration grudge given the case spanned two administrations.
While accessiBe may be silent about this on its site, in January it was already framing the settlement in FTC Fines Israel’s AI Web Accessibility Company AccessiBe $1Million Over Deceptive Claims by Jewish Business News, which devoted half of the article to accessiBe’s messaging. In the article, accessiBe claims the FTC complaint only focuses on 2019 through 2022 (which suggests accessiBe knew it was lying for at least three years) and deficient testing (which is not human impact).
Of course, I have documented accessiBe doing these things both later than 2022 and with documented tests. That said, the FTC’s post provides a good rubric for watching accessiBe and its almost guaranteed future transgressions:
To resolve the allegations of the complaint, accessiBe agreed to tell the truth in its ads, disclose connections between endorsers and the company’s products that might affect the weight consumers give the endorsement, and pay the FTC $1 million to provide refunds to consumers.
I don’t know (yet) how to report when accessiBe violates the terms of the settlement, but for now when you see it maybe document it and drop a comment here.
It’s a post reproduced almost exactly from the accessiBe blog. The byline is the content aggregator, Stacker, but the post does reference the original accessiBe author. Interestingly, after I left a comment on the article referencing the FTC case the morning of 20 April (7:54 AM), the post disappeared from the site completely.
Thank you, Adrian, for covering this in such detail. I plan to cover this in my next podcast.
I am not surprised by this news because of accessiBe’s claims. I am, however, surprised by the settlement, which seems extremely low based on their revenue and the entire nature of their business making these claims.
Thanks, Adrian. Are there any accessibility services that you DO recommend? I run a very small non-profit that publishes a multimedia magazine. We have used accessiBe, but will now look for other solutions. But it is not practical, feasible, or affordable to do a customized fix–for what we continually publish, and even more impossibly, for our vast, 10-year archive. There are obviously many other websites in similar predicaments. What do you suggest?
This is an important step from the FTC, and credit to Adrian for years of consistent scrutiny in this space.
We agree that overpromising harms trust — not just in vendors, but in the broader push for inclusion. At AAAnow, we’ve argued that accessibility overlays aren’t a silver bullet — but when used transparently and responsibly, they can be part of a broader, practical strategy.
Rather than take sides in a binary debate, we’ve tried to offer a more balanced perspective on where these tools can fit in. You can read our take here:
Thanks, Adrian. As ever, clear, thorough, and unapologetically honest.
At AAAnow, we’ve been pushing for a more realistic conversation around accessibility tech. The FTC’s action may feel overdue, but it’s a vital marker for the industry.
We recently published a piece reflecting on why some of these overlay tools gained traction in the first place, and how we can move the debate forward without ignoring the harm.
Spoiler: it’s not just about the tools themselves, but how they’re pitched, purchased, and misused.
Dear AAAnow Sales Team (I say that because the email address you used is your sales email account),
I appreciate that you felt so strongly about promoting your post that you did it twice. On a post that demonstrates how lying about conformance and “AI” use can get the FTC after you. It’s a bold move.
I also appreciate that your domain name redirects to LinkedIn instead of a real site. Though that may be a function of only having three employees, two of whom seem to be carrying the similarly creatively-named AAATraq into this new venture (which should be great for ranking in phone book listings).
As you requested, I’m going to share my thoughts on your piece here and also on the LinkedIn post itself.
It’s time for a more balanced, honest, and perhaps even ‘unbiased’ conversation.
I don’t know why you put “unbiased” in scare quotes, but if your intent was to be ironic I think it works here.
You outline how overlays can help users:
Contrast controls for older users or low vision
Dyslexia-friendly fonts and reading masks
Focus assist for ADHD
Text enlargement and simplified layouts
Real-time colour-blind adjustments
Support for cognitive load reduction (e.g., turning off animations)
This presumes a user understands these controls exist, can access the controls, and that the features work.
But this also presumes the user doesn’t already use their browser or operating system to do the bulk of these. For example, contrast features are built into operating systems (for all users, not just older or low-vis users). People can scale text in their browser already. Folks can use reading mode for a simpler layout. Color blindness adjustments are typically done with contrast settings, which, again, the OS provides.
I agree browsers don’t support dyslexia-friendly fonts by default, though evidence in support of those is scant. I also agree they don’t provide “focus assist,” though data on that is also scant (please link peer-reviewed material if you have it). As for disabling animations, I have yet to see an overlay successfully stop animations even on their own sites.
These are real, immediate improvements – delivered at scale, without waiting months for development updates or a redesign. For many organisations, especially those without enterprise budgets, these tools are a game-changer.
But they aren’t real. And less immediate than browser or OS settings. The ones not provided by the OS aren’t demonstrably useful for enough of the audience when balanced against all the risks — to organization and users — of overlays.
Talk to people who use these tools regularly, and one thing is clear: they’re not just a nice-to-have. They’re essential.
If you mean overlays by these tools, I do talk to users regularly. I’ve spoken to them for years. They’ve been vocal about how overlays introduce problems for them (there are even news articles!).
For users with ADHD or dyslexia, overlays that reduce distractions or adjust typography can mean the difference between comprehension and confusion.
You are cherry-picking the audience for whom the needs and outcomes are so incredibly varied that there is little or no peer-reviewed data showing these tools (Dyslexia-specific fonts and focus assist), whether enacted by overlays or not, broadly help.
For older adults or users with visual impairments, being able to magnify, adjust spacing, or trigger high-contrast modes can mean the ability to complete a task without frustration or giving up.
Well, with the exception of adjusting spacing, the operating system provides those. But that tends to align more with dyslexic users.
The accessibility purist view is that only manual audits and code-level fixes provide true compliance. And in technical terms, they’re not wrong – widgets can’t fix poor HTML, broken ARIA roles, or complex forms.
One the one hand, that’s not a purist view. Programmatic exposure to platform accessibility APIs can only be addressed in code. On the other, manual reviews are just one item in the toolbox, especially when dealing with the varied needs of users across all the contexts of the the things they encounter.
But there’s something else worth considering: do the many critics have a position to protect?
Manual accessibility work – audits, dev remediation, consulting – is expensive and time-consuming. Entire businesses are built on selling these services. Automation, overlays, expedience platforms, and accessibility-as-a-service models disrupt that market.
Could some of the criticism stem from bias? It’s worth asking.
There really isn’t a position to protect. Digital accessibility work pays poorly and leads to burn-out. People aren’t doing it to get rich. I also know none who do exclusively manual audits nor even want to. It’s time consuming and frustrating, which may be why you don’t see companies doing it (other than some boutique shops). Heck, I don’t do it.
It’s easy to call overlays a ‘shortcut’. But what if they’re actually a way to start? To do something rather than nothing? To improve experiences for real users, right now, instead of waiting on budget cycles and development sprints?
Well, I never called them a shortcut. I mostly call them not fit for purpose.
The future of accessibility isn’t widget vs. manual. It’s widget plus manual. It’s compliance plus usability. It’s not just what the code says, it’s what the user experiences.
Yet we continue to establish that overlays, broadly, do not work and increase risk. Nothing you have offered comes close to refuting years of data, articles, research, legal cases, and more.
Now that I’ve reached the end of the piece, it’s not clear what you’re prepared to do. Other than carry water for overlay vendors, obliquely by framing it as an appeal to purchasers (and not disabled users).
Let’s have a real conversation. This space needs less chasing the impossible and more pragmatism.
I have this conversation constantly. There are millions of dollars of venture capital forcing me to have this conversation on a weekly basis. I have been sued for having this conversation. I’m having yet another conversation right now, with a sales team on my own blog.
So what, as an organization, are you going to do to help disabled people as a result of this conversation?
Thanks for taking the time to respond, and for highlighting the strength of feeling this topic still generates.
We stand by the article’s core message: this debate needs more balance. It’s entirely fair to critique false claims and poor implementations — and we don’t dispute the technical limitations of some products. But to insist that any technology in this category is inherently harmful or misleading ignores a more important truth:
Not all website owners are technical. Not all organisations can afford manual audits. But many users still need support… today.
To be clear, we’re not saying overlays solve everything. They don’t.
We’re saying that technology can, and should, play a role in improving accessibility for real users, especially when the alternative is doing nothing.
You make the point that operating systems and browsers already offer tools. That’s true, but awareness, access, and usability of those tools remain barriers for many. A clearly visible panel on the page is sometimes the only way a user even realises support is available.
Your argument also seems to assume that “overlay” equals “non-functional.” That might be valid for older implementations or vendors making overblown claims, but it dismisses the real and growing field of UI personalisation tools, AI-supported contrast correction, and experience panels that are evolving rapidly.
This includes features like:
– Real-time color contrast control (still the #1 WCAG issue, year after year per WebAIM)
– Reading modes for neurodivergent users
– Simplified views for cognitive accessibility
Should these things be fixed in code? Yes.
Will that happen for every SME, e-commerce template, or legacy CMS? No.
So the pragmatic question is: why not use technology to help some users now, while continuing to work toward better standards, education, and long-term fixes?
Dismissing everything under the umbrella of overlays feels like gatekeeping — especially when done by those who sell manual services. We’re not saying that’s your motivation, but it’s fair to observe that there’s a professional interest at play across the whole sector, on both sides.
So let’s return to the original ask: a balanced, open conversation. Not marketing spin. Not purist dogma. Just a recognition that progress isn’t binary, and that technology has a real role to play if implemented transparently, ethically, and as part of a broader strategy.
You ask what we’re going to do. Here’s the answer:
We’ll keep building tools that support real users, in real time, without waiting for perfect conditions.
You don’t have to agree with our approach – but we hope we can at least agree that doing something useful now is better than waiting for perfection that may never come.
We stand by the article’s core message: this debate needs more balance. It’s entirely fair to critique false claims and poor implementations — and we don’t dispute the technical limitations of some products. But to insist that any technology in this category is inherently harmful or misleading ignores a more important truth
I am happy to debate / discuss something when there is new information. But you are essentially asking people to restart a mostly settled discussion without bringing new information.
Broadly, the industry doesn’t make a blanket statement about any technology in this category. What you are actually hearing is acknowledgment that the best predictor of future behavior is prior behavior. Heck, it’s even current behavior as I wrote yesterday.
Not all website owners are technical. Not all organisations can afford manual audits. But many users still need support… today.
Then instead of regurgitating the same talking points, take action. Take the lessons we have all learned and make something better. Outline what an ethical overlay would look like. Identify how to educate users about existing features in their platforms. Get some funding for follow-up research. Really, anything other than trying to promote a business model shown to rely on FUD at the expense of users. It has plenty of venture funding, it doesn’t need more boosterism.
Your argument also seems to assume that “overlay” equals “non-functional.” That might be valid for older implementations or vendors making overblown claims, but it dismisses the real and growing field of UI personalisation tools, AI-supported contrast correction, and experience panels that are evolving rapidly.
This includes features like:
Real-time color contrast control (still the #1 WCAG issue, year after year per WebAIM)
Reading modes for neurodivergent users
Simplified views for cognitive accessibility
You are falling back to two audiences for whom the needs and outcomes are so incredibly varied that there is little or no peer-reviewed data showing these tools (Dyslexia-specific fonts and focus assist), whether enacted by overlays or not, broadly help.
Contrast is indeed a challenge, but every overlay tool I have seen leans on CSS to achieve its goals, and is often less effective than features built into the OS or browser.
But if you broadly want to rebrand overlays as personalization tools (which many are doing to avoid false claims lawsuits and class action suits from their own customers), go for it. But your “discussion” (scare quotes) needs to be frank about the barriers and legal risk those overlay, sorry, personalization products introduce today. Barriers to users, legal risks to customers.
Somehow your conversation keeps avoiding the very real facts that a few non-specific and questionable features bring a far greater pile of problems. It fails to warn that any solution has to address those before claiming it’s better than nothing.
So the pragmatic question is: why not use technology to help some users now, while continuing to work toward better standards, education, and long-term fixes?
Dismissing everything under the umbrella of overlays feels like gatekeeping — especially when done by those who sell manual services. We’re not saying that’s your motivation, but it’s fair to observe that there’s a professional interest at play across the whole sector, on both sides.
The question and statement make assumptions that technology is neutral. Overlay makers are anything but. They have relied on lies and dysfunctional technology. It’s not gatekeeping to be wary of creating more harms.
However, I agree there is a professional interest here.
Overlay companies have been overwhelmingly well funded. They have taken in tens of millions (I think hundreds now) each. One even overtook a traditional auditing company, installing its leadership at the top. Overlay companies have spent money repeatedly suing critics around the world, in the ultimate act of gatekeeping.
So yeah, there is a professional interest here and it’s about making that sweet investor money to lock up customers for predictable, repeatable revenue while paying shills to astroturf their work.
So let’s return to the original ask: a balanced, open conversation. Not marketing spin. Not purist dogma. Just a recognition that progress isn’t binary, and that technology has a real role to play if implemented transparently, ethically, and as part of a broader strategy.
Well, as far as I know, we’ve had the open conversation and you seem to dismiss the reality of the last ten years of behavior by overlay companies (and their technical failings). It’s an ethical gap to dismiss that, in my opinion, and no conversation is complete without discussing an ethical framework at the same level as blue-skying the technology.
As for transparency, what is your revenue stream?
You ask what we’re going to do. Here’s the answer:
We’ll keep building tools that support real users, in real time, without waiting for perfect conditions.
Cool. Take the lessons learned (and the research I cited) and consider building a browser extension to cover the gaps you keep naming. An extension can handle the font swapping, contrast changes, and text highlighting using existing technologies. An extension also doesn’t rely on enterprises and their budgets to deploy; no marketing to convince companies (whether by FUD or appealing to their good nature).
An extension also can retain user settings across sites. So a user can benefit regardless of web site. Those settings will also be sandboxed, providing greater security and privacy, instead of relying on vendors following GDPR. Users could export settings, add customizations, share preferences, and so much more. And the cost to users would, or could, be absolutely nothing.
I understand there’s no revenue model there, but if you want to keep building tools that support real users, in real time, without waiting for perfect conditions then I think that approach offers the greatest reach with the least risk of harm.
Here’s your opportunity to show you mean what you say.
It’s cool and good and great to see the AAAnow overlay team trying to blend crisis response and SEO boosting into the same effort, without actually delivering anything substantive counter claim-wise.
Intensely frustrating to read comments by AAAnow that look entirely LLM-generated while Adrian wastes his precious human time and brain cycles debunking such anodyne low-effort slop. Though I suppose that’s the point. The contempt for the accessibility profession and the disabled people served by it couldn’t be clearer.
Leave a Comment or Response