Making Digital Accessibility Work In The AI Era | Dr. Michele A. Williams

The AI Governance Briefing

Episode: Making Digital Accessibility Work In The AI Era

Host: Dr. Tuboise Floyd

Guest: Dr. Michele A. Williams

,: 2026

Cleaned transcript — lightly edited for readability. Speaker labels and timestamps preserved from original recording.

================================================================================

Dr. Tuboise Floyd

:: 00:00

Welcome to Human Signal. 97% of the web still presents accessibility barriers to disabled people. That's not a typo. That's our digital infrastructure.

Pages average over 50 preventable errors each. Missing alt text, broken keyboard paths, unreachable forms.

If you're a Gen X leader running a team, a product line, or a whole division, this isn't an edge case. This is your user base, your legal risk, and your culture baked into every screen you ship. I'm Dr. Tuboise Floyd and this is Human Signal.

Authentic intelligence in the digital world. Today we're asking: how do you make digital accessibility actually work when AI is touching everything from UX research to captions to code?

And I'm not asking in theory. I'm asking someone who lives it.

My guest today, Dr. Michele A. Williams — Accessibility Consultant, UX Researcher, and author of Accessible UX Research, published by Smashing Magazine.

Michelle helps product, research, and policy teams rethink how they design, test, and ship so that disabled people aren't an afterthought, but a design constraint from day one. She's the person organizations call when they realize they never actually did accessibility in the first place.

Michelle, welcome to Human Signal.

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 01:53

Thank you so much, and thank you for having me. I'm glad we're talking about accessibility and where it actually lives in real teams and real organizations. And of course, with AI changing the landscape every day, let's start with leaders.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 02:09

Who are quietly anxious right now. The Gen X product VP, the engineering director, the civic tech lead. They see AI everywhere. They see lawsuits in the headlines and know accessibility is important, but it just feels like a compliance fire. Michelle, from your vantage point, what's really at stake for them?

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 02:35

So I'd say let's look at it in maybe three buckets. The first is equity — who is getting to use your systems, and who are you either quietly or really loudly shutting out? The second is resilience — whether your products handle real-world diversity or they fall apart when you have what you believe to be edge cases. And then the third would be trust. When people repeatedly experience access barriers, they leave. They lose confidence in your brand and your institution and your leadership. Not having accessibility in these areas is already doing that currently.

AI is then magnifying that because now you are moving more rapidly. But you're not necessarily catching those problems — you're magnifying them, increasing them, encoding them, and accelerating them at scale. So this is what we need to be paying attention to.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 03:37

Wow. So the risk isn't just we missed an alt tag — but rather we automated discrimination into our everyday operations.

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 03:49

Yeah, exactly. And because AI is so fast and almost invisible, in some ways it just amplifies what's already been happening — that accessibility has primarily been missed.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 04:01

In your book, you start with disability mindset. Before we get tactical, we have to talk about how people are thinking — or not thinking — about disability. What's the mindset problem inside teams?

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 04:20

What I see inside teams is really reflective of the larger society. And when I say society, I'm not just talking about the United States — we see this worldwide. Disability is often othered and made invisible. It's the largest minority group, and yet it's still seen as a minority group that doesn't need to be accounted for.

That's because non-disabled folks often feel that disabled people are not full contributors to the world. This is a concept known as ableism — discrimination against disabled folks based on the assumption that non-disabled ways of moving, thinking, and communicating are the default, and everything else is a deviation and even to some extent lesser.

This relates to what are known as models of disability — the ways that we think about disability and the actions that result from that thinking. There are several models, but two in particular: medical versus social. The medical model primarily says disability is something to rehabilitate or institutionalize. The social model flips that — it's not the disability that's disabling. It's the lack of access and the negative thinking by non-disabled folks.

So in the working world, if people in positions of power have that medical model mindset — which is often the case — then disabled people aren't seen as equals, their needs aren't considered, and their inclusion in the process of building isn't either. Until there's a mindset shift, the actions won't truly change.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 05:57

In your book, you share a story about UX researchers who asked: if we have accessibility specialists, why do we need disabled people in our research? What's broken in that question?

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 06:17

That was experienced UX professionals on a call with folks they didn't even know had disabilities — and they said that. One of the responses I had: we have design guidelines, but we don't say "why do we need to talk to users?" — meaning non-disabled users. We know that guidelines are not enough.

But somehow, when we put disability in the mix, people are strangely comfortable substituting checklists for lived experience, if they get to it at all. The deeper issue is that we don't yet see disabled users as just users — as part of the user base. They're treated as optional, not yet considered.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 07:10

There's a language discussion in your book about the dichotomy — how disabled people reference themselves versus how non-disabled people reference disabled people.

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 07:26

There are currently three primary ways that we talk about disabilities.

You're probably most familiar with person first — "person with a disability." That's how folks are typically taught. It came from the cognitive impairment community — specifically from folks with intellectual disabilities who said: see the person before the disability. And to some extent, that is accurate. But what it ended up doing, unfortunately, is making disability seem like a bad word, a taboo word.

Then there are activists and many disability groups across many cultures who have decided that disability is just a descriptor — so why not put it first? That's known as identity first. That's how I tend to align — disability is not a bad word. It's just part of the diversity of the human experience.

The third way is usually to avoid disability language entirely — say "handicapable" or something. I've rarely seen actually disabled folks prefer that language. It doesn't feel as empowering as just being able to say the word.

The idea is that if we normalize it enough — understand it's just part of our diverse and beautiful human experience — it doesn't need to be a big thing. And we don't want language to stop us from doing the actual work.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 09:20

Just hearing you say "handicapable," I cringed. I'm from Alabama. If you ever hear "God bless your heart," that's an insult — not a compliment. So when I hear "handicapable" — I cringe. That's just where I am on it now.

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 09:45

What people are trying to say is: take the "dis" out of disability, focus on the ability. Some of those terms go back to person first in spirit. I won't say they all give you as much of a cringe.

But this goes back to ableism. Disability as a whole is seen as a negative — so someone has to do something to reclaim their power. Words and language become a way to establish: I don't align with this idea that this is negative. I align with the idea that this is okay.

But ultimately, it's about respecting individuals — and really about changing the disability mindset. If we had a different disability mindset, folks wouldn't have to work as hard to find the right language. Because what they're really working against is discrimination.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 10:48

I really like the approach where you emphasize the individual first. In practice — when I'm referring to a person, could I say "a person with a vision impairment?" Is that correct?

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 11:07

That is one accepted way. But most people I know with vision impairments say: just call me blind, or I have low vision. This doesn't need to be a thing.

I'd say: whichever feels most comfortable for you — person first or identity first — start with that and then confirm it with the individuals you're talking with. If you're writing an article, run it by people with more understanding in that area. Start to find your comfort zone, while being respectful of the community you're most engaged with at the time.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 11:47

So if I ever get stuck and I'm writing something, I'm calling you directly, Dr. Williams, to help me clarify.

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 11:53

Yes. But also remember — the number one preferred way a person wants to be identified is by their name.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 12:00

Correct.

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 12:01

That's why I say: at some point, don't get too hung up on the language. We're all just part of the human race. Language matters — but I think people get too hung up on it and don't do the other work, which is more important.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 12:16

I try to learn people's names. I have a non-standard name myself — when people mispronounce it, we have a good laugh and move forward. It's always about putting the human into the interaction. Putting the person first.

I'm going to challenge the Human Signal listeners: do as Dr. Michele A. Williams recommends. Put the person first. Ask the individual. If not, reach out to experts in the field.

Let's touch briefly on AI. Where does AI sit for you in this picture of disability?

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 13:08

AI has obvious potential to do great things for accessibility — particularly at scale. But only if the current tactics are disrupted. We also see that AI has already done real harm to the disabled community — in hiring algorithms, benefits decisions, surveillance technologies. Yes, AI is great at some tasks the accessibility community has championed — like captions and alt text. But ultimately, AI is a tool and not a savior.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 13:45

You bring up an excellent point. And in your book you talk about explicit language choices — "disabled people," identity first. Why?

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 14:08

There was literally a campaign by Easter Seals with the tagline "disability isn't a dirty word." Many communities — Deaf, autistic, blind — see disability as just the core of who they are. It's not something to be hidden behind.

I'll note: disability and impairment are not always how some groups identify. The capital-D Deaf community, in particular — deafness is not seen as a disability or impairment. It is simply an identifier. That's the cultural model of disability. The capital-D Deaf community is a culture. Deafness is not necessarily associated with disability or impairment in that context.

Nonetheless, what's most important is to ask individuals what they prefer and honor that. And don't let fear of perfect wording stop you from building relationships and doing the work.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 15:23

The social model of disability is one that really jumped out at me in your book. The primary idea is that disability is caused by environmental barriers and social construction — not the individual's impairment. My light bulb went off. Sometimes at family functions it's too loud and I just tune out — that's my way of relating to the concept. Is that on the right track?

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 16:22

Yes — what that's saying is that certain environments do and don't work for you. If I'm consulting on a physical event, I'll say: you need a quiet area and a main area, because some folks need a decompression space away from the crowd. That's a known adjustment across many conditions.

That is changing the environment — taking the accountability upon the planners to acknowledge the diversity of the people attending. Too often we put it on the person to make all adjustments needed to participate, when there are just as easily opportunities for us to make those adjustments ourselves. A portion of that is acknowledging: if we just made the environment more inclusive, maybe disability wouldn't feel as disabling.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 17:41

As we pivot: there's a simple shift I keep seeing, and it occurred in me reading your book. It's designing with disabled people — and that's as much about power as it is about process. You agree?

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 18:26

Yes — and. You should be including disabled people in your processes. And this includes recognizing disabled folks as peers and leaders, not just as recipients of what you make.

I'll note: I don't use the phrase "design with, not for" — because I think that phrasing still positions non-disabled people as the authority. When someone says "design with, not for," they're usually still envisioning a non-disabled designer including disabled folks. But really, disability is part of the diversity of all human groups. Disabled people should be showing up in the defining of the problem, the shaping of research questions, the evaluation of solutions — all throughout, with diverse groups of people working on them.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 19:32

Say that phrase again — I want to make sure I've got it.

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 19:39

"Design with, not for" — that's what you hear people say. But I don't use it. It still positions non-disabled people as the authority. There's another phrase we'll get to later that I like better.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 19:52

Let's zoom out to the system. Many teams hear accessibility and immediately picture a blind person using a screen reader. That's part of the story — only a small part. Can you sketch the broader landscape?

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 20:21

People tend to place disability in one or two categories. In physical spaces, only wheelchair users. In digital spaces, only blind users. But there are dozens of disabilities with wide-ranging impacts.

In digital tech, we tend to use five major categories: vision, hearing, mobility, speech, and cognition. Within these are varying disabilities and degrees of disability.

Within vision, for example: more people are low vision than fully blind. Low vision folks may hold objects close or enlarge them. They use magnification software. Color contrast — making sure text and background are sufficiently different — addresses a need for low vision users.

A deaf participant in my research flagged that a client was using phone calls as the primary communication method — which creates an immediate barrier for deaf or hard-of-hearing users, as well as for people who don't use voice.

The categories are a way to be systematic in our approach, but they don't represent how humans actually show up in the world. And we still have a lot to learn about the cognitive categories — any way that people think and process information. All of this shapes how someone experiences your UI and your AI-driven content.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 22:34

You can't assume anymore. You can't lead with old mindsets when you're designing products for use. All things must be considered. I know we can get into the minimum viable product argument, speed to market, first mover — but you talked earlier about putting the person first. That's the same overlap with Human Signal. I say: put the human first in everything we do and say.

I encourage our listeners to get your book, Accessible UX Research — dive into this chapter, get the nomenclature, broaden your viewpoint on the multifacetedness of the human race.

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 23:55

The only thing I'd add, going back to the social model: sometimes we wouldn't have to ask if the environment was just accessible. If you built products in a way that included established considerations and best practices for the diversity of your users, we don't have to stop, disrupt, ask, fix, and do jarring cycles. We could just be doing the right things for the people we're doing them for — right from the start.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 24:44

And we're going to get to an action plan — because that's what we advertised and listeners have been DMing me about. Dr. Williams is going to talk about the action plan. Stay tuned.

But first: what does exclusion in UX research and code actually look like on a real team?

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 25:17

A couple of common patterns. First: representation. Teams will say "we don't have disabled users for our product." But disability is much broader than what we're thinking about. It's statistically near-impossible to have a group without disability. The CDC's numbers put it at about 1 in 4 U.S. adults having some form of disability — and that's still underreported because diagnoses for cognitive categories keep improving.

When teams say "we don't have disabled users" — what they mean is: we're not thinking about disabled users, or we're leaving them out to prioritize something else.

Then there are the tools. Oftentimes the things we're using and the things we're making are part of what I call an inaccessible digital ecosystem. Even if we wanted to include disabled folks, sometimes they can't participate in the workforce or the products we're building because the tools themselves present barriers.

And when we talk about AI-generated code and content: it was built on what you said at the very beginning. 97% of the web has accessibility issues. If AI was trained on what's present, and what's present is inaccessible — what's produced is inaccessible as well.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 26:52

Generated code and content, out of the box, is trained with standard models — and those standard models are not disability-first by design. Products get shipped that look fine visually but don't have the correct semantic structure. No labels. No logical focus order. That's chaos to assistive technology.

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 27:24

Literally — that's what defines inaccessibility. For your listeners: a lot of inaccessibility comes in the web and mobile app domains because of how things get coded.

There are ways to code that are semantic — meaningful — that assistive technologies can understand. Assistive technologies are the devices that people with disabilities use to navigate and accomplish computer-based tasks.

If I'm using a different setting than the default — say I need larger font, recolored display, or text read aloud — in order for those things to work, the assistive technology needs to understand the code. It needs to know how to read content aloud, how to engage with a keyboard versus tapping.

When we say something is not accessible, it's because the code is not giving the assistive technology the information it needs. Technology has moved so quickly that there's always a lag in making new innovations accessible. Standards come after innovation. And by the time something catches up, it's proliferated. We now have significant tech debt in terms of inaccessibility.

When you build tools that allow anyone to make a website or an app — without requiring accessibility considerations — and then AI is built on that foundation — that's where we are.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 29:11

Let me give you an example. One day I was rushing to get to a meeting. I said to Siri: "open Zoom." Siri didn't open the Zoom app. It opened the device zoom configuration and physically magnified my entire screen to maximum. I had never experienced that setting before. I panicked — I thought I was going to miss a meeting with the president. My phone was locked. I had to go home, sit at a computer, Google how to get out of that setting.

That was my experience with assistive technology on an iPhone — Siri misinterpreted my voice command and triggered an accessibility feature I didn't know existed.

So I can only imagine what our disabled friends go through every day.

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 30:21

One thing I'll pull from that: get familiar with what your devices have built in. A lot of settings on your devices would benefit people even if they don't identify as disabled. Anyone over 40 can benefit from larger text. Period.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 30:37

Hello.

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 30:38

There's been a lot of talk about putting all these settings under a category called "Accessibility" on your phone — when really they should just be part of standard settings. And unfortunately, Apple named their magnification feature "Zoom." I've noticed that too.

But if we had normalized the idea that your phone can magnify — that it's just a setting, not a special accessibility thing — you wouldn't have panicked. You would have known how to undo it.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 31:05

Exclusion becomes the default setting — not because anyone chose it, but because no one designed against it.

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 31:15

The hope is that we are more intentional not to do that. But by and large — exclusion is the default setting. Yes.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 31:23

In your book, you write about disability dongles — solutions that look clever to non-disabled people but don't actually help disabled people. Give us an example.

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 31:36

I'll first acknowledge disability advocate Liz Jackson for coining this term. A classic example: the smart cane. We know that blind pedestrians use a white cane. Engineers will often add sensors, vibrations, or connectivity to the cane. But when you talk to white cane users, you hear why these devices really aren't meeting the need. They're heavy. They're expensive. They're distracting. They don't reliably tell you where an obstacle is, what it is, or what to do about it.

So they're not solving a real problem. And if you dig deeper, what many blind pedestrians actually want is safer environments — not more technology on their cane.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 32:17

What about something like the Meta glasses for navigation — where does that sit for you?

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 32:22

The Meta glasses are beautiful. But — be careful with privacy and all of that. Every tool has its caveats. I didn't say Meta gets a gold star — I said people use it. There's a difference.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 32:44

Fair. To exit this section: the real move in design isn't more tech for the disabled person. Your book is saying: remove barriers from the system.

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 33:04

It could be both. If we're aligned with the social model — yes, before we ask "what can we build," let's be clear about whether we need to build something and what a meaningful solution actually looks like. But if the solution is technology, then we make sure that technology doesn't present yet another barrier in our efforts to remove a barrier.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 33:36

Let's come back to AI. Leaders are being sold that AI is the miracle cure for accessibility — we'll just auto-caption everything, auto-generate alt text, and auto-test our UI. Is that generally helpful? Where does it backfire?

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 34:02

Auto-captions are a big step from no captions — especially if you have a large bank of media. Alt text automation helps get the task moving. Automated tools are actually important for flagging issues at scale, if taught to do that correctly.

But where it backfires: auto-captions are often not good enough. They need correcting to be fully comprehensive. Alternative text is not always the right description — a woman in a meeting room doesn't necessarily tell you why an image has been included in a story about women in the C-suite. And automated tests, even when programmed correctly, only cover a portion of WCAG guidelines. Even if accurate — which is a big if — they still may miss the overall user experience that only manual testing captures.

AccessiBe is a company that often comes up in this area. Look at the million-dollar fine they're facing for false claims they made about their platform's capabilities. That will give you some understanding of hype versus reality.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 35:32

Penalties and fines. Wow.

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 35:36

And backlash and lawsuits. Be careful.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 35:41

So the pattern is: AI can draft, but the human needs to review. And that review has to include disability expertise — not just a quick glance. And you can't use AI as a stand-in for that expertise.

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 36:01

Oh yes. We have to talk to real people. Real human in the loop — not the imitated human. AI should be a tool embedded in a thoughtful, accessible process — but not a replacement for disabled participants or human judgment.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 36:26

Absolutely. Because with AI, it will hallucinate. It will tell you with great confidence that it is right when it's wrong. That's why you need subject matter experts to walk through and review.

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 36:44

There's some basic knowledge that all of us should have been taught — in every craft and every field — some manner of disability inclusion. It's not to say you need an expert for everything. The point of an expert is the more complex part of whatever that is. There should be a baseline understanding. That's why I wrote the book and do trainings — my mission is to make accessibility accessible. It should not be a topic people feel uncomfortable with. We all should have some basics. Then you bring in expertise.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 37:23

A lot of leaders are stuck at: "we passed automated checks — what more do you want from us?" When organizations bring you in, what are they usually stuck on, and how do you help them move beyond the checklist?

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 37:49

First, if someone has even passed an automated check, that's doing great — usually teams aren't even aware to run that check, or they're not passing it.

But regardless of where you are, there are generally sticking points. One is what I'd call a compliance plateau — maybe you've had training or run a scan, but nothing is really improving for real users. Accessibility gets talked about in the abstract but never actually makes it into the product.

Then there's tool mismatch — the tools you're using don't allow disabled teammates to fully engage, or don't support disabled research participants.

And then leadership. Leadership has to advocate for this work to get done. That may look like helping them understand what this is about — and translating that into policy, procurement, and roadmaps. This is how we shift from "did we check the box" to building it into everyday processes.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 39:24

What I'm seeing is: you are not just doing audits. You are redesigning how teams work.

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 39:41

Because an audit is just a snapshot. I'm more interested in helping teams build a practice that will sustain after I'm gone. I don't want you to need me forever. I want you to feel confident this is how you operate.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 39:57

We promised listeners they'd walk away ready to commit to 90-day accessibility practices their teams can sustain. Let's make that concrete. If I'm a Gen X leader in a real organization, what does a 90-day arc actually look like?

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 40:41

Let's start with the first 30 days — establishing your baseline.

Scan where AI and accessibility are already intersecting in your organization. Where are you running auto-captions, auto-summaries, auto-generating UI, copy, or code? For each of those places — what external, non-AI tool or resource can you use to validate what's being output? An internal team review? An automated scan tool based on established accessibility criteria?

Based on that scan, you can start to see where you currently stand in your ability to produce accessible outputs.

Then ask your research or product teams: are we including disabled folks in our studies? In our outreach to consumers? If the answer is no, or "I don't know," or "never" — flag that.

The goal of the first month is to stop guessing and learn how your system is actually behaving.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 42:12

So month one is reality testing — not heroics.

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 42:17

Absolutely. Start with knowing where you are — then map out what's next.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 42:23

Excellent. Now we're moving into days 31 through 60.

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 42:30

Day 31 through 60: start to change those defaults. Begin disrupting the patterns you identified.

Is your design system built accessible? Are your coding tools overriding accessibility? Are your teams equipped with the expertise to know when this is occurring?

Wherever there are gaps — understand what will help overcome them. Is it training? Different tools? Hiring practices?

Then add accessibility and AI impact questions to your procurement. Start asking vendors: does your platform meet accessibility standards? Do you have documentation to show that? Do your AI tools allow human-in-the-loop review of outputs to ensure compliance?

And: choose an upcoming research project and make inclusion of disabled participants non-negotiable. Build that time and budget in. This is where you stop relying on individual champions or grassroots efforts and start putting this in your infrastructure.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 43:54

And once it's in procurement and templates, it's much harder to quietly roll back.

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 44:02

That's the hope. You're building in guardrails — not just giving pep talks or talking about it in the abstract.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 44:09

Guardrails are in place. Not giving pep talks. Now we're in the last 30 days. Talk us through that.

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 44:24

I'm trying to be realistic — disruption is not easy. We don't want everything solved by day 90, but we can start making real steps toward normalizing this.

Actually run that inclusive research you planned. One thing that may be difficult: finding participants who are disabled. But even in that difficulty — start to understand why it's difficult, and then start leveraging disability-focused networks to break that cycle. Make sure you're using accessible tooling to capture feedback.

Then, because you have that research done — bring those clips to leadership reviews. Let the executives hear how folks are navigating your products with assistive technology, or how AI-generated auto-captions don't work. This starts to normalize the idea that this is not okay, that we're going to do things differently.

Document what you're going to change as a result. You've probably learned something — from scans, from talking to users. Write it all down. What are we going to fix? What needs to be postponed? What doesn't ship because we haven't made this change yet by day 90?

Accessibility isn't a special project we're going to get to. It's part of a feedback loop. Part of building that muscle of thinking about it, asking about it — and starting to feel normal.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 46:15

And this is where Human Signal intersects with your work. Our mission is to help real operators slow and safe rollouts, rewrite procurement criteria, and build governance around reality instead of vendor hype. Accessibility is one of the sharpest reality checks we have.

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 46:36

If your AI strategy is producing systems that disabled people can't use, that's a pretty strong sign your governance is performative and not protective.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 46:47

Your final chapter zooms out to disability inside the UX research profession itself. What do you want leaders to hear on that front?

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 46:57

Two things. First: recognize disabled professionals — not just disabled participants. Don't just consult disabled folks occasionally. They should be hired, promoted, and be real decision-makers. Statistically, you likely already have disabled leaders in your organization, whether they identify or not. But normalizing diversity in these spaces must be inclusive of disability. If your tools and workflows are inaccessible, you're literally locking out disabled talent from the field and from leadership. I would really implore you to make sure you are not creating those barriers.

And zooming out further: AI policies, safety frameworks, research standards — all of these need to be built with disabled folks in the room, so that we're not rebuilding the same exclusion at higher speed.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 47:59

So if we're serious about AI governance, we must include disabled people — and we must have them be among the people we design with and answer to, as a built-in accountability.

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 48:18

Yes. And let me give you a phrase to put in your head.

Historically, the disability rights movement gave us the mantra: nothing about us without us. People were making decisions about disabled folks without consulting disabled folks. So the mantra became: nothing about us without us.

But it has evolved. Nothing about us without all of us. There should not be spaces where someone is left out intentionally. That goes back to the mindset — our mindset informs our actions. Nothing about us without all of us.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 49:03

Michelle, thank you for your leadership in that — because you held me accountable with all love and with the utmost kindness, and I'm very appreciative of that.

For the leader who's listening and feels both convicted and overwhelmed — what's one action you want them to take this week?

--------------------------------------------------------------------------------

Dr. Michele A. Williams

:: 49:28

I don't want people to walk away thinking I'm not living in reality. The reality is that including accessibility into your pipelines might be a significant shift — particularly if it hasn't been explicit to this point. But each day brings more legal risk when you're not doing it. Delaying is not the answer.

Start with probes — like we talked about earlier. Just ask around. Ask if people in your organization are even familiar with this concept. There may be people currently doing it who just aren't formally empowered to say so — it's not in their job description. Someone may have an interest in starting, they just haven't been empowered. Teams often have in-house champions they aren't fully leveraging.

And sometimes, good build practices are unknowingly leading to accessible outcomes — because good user experience, good code practices, and inclusive policies are the foundations of accessibility. The main thing is to start asking. Start probing. See where you are.

--------------------------------------------------------------------------------

Dr. Tuboise Floyd

:: 50:34

If you're thinking you need help — this is literally what Dr. Michele A. Williams does. She works with organizations to redesign their research practices, product decisions, and workflows so that accessibility is built in, not bolted on.

To go deeper: pick up her book, Accessible UX Research, by Dr. Michele A. Williams, and visit her online to learn how to bring her into your team as a strategic partner. You'll find links to the book and to Michelle's work in the show notes.

Subscribe to Human Signal — Authentic Intelligence in the Digital World — on Spotify, Apple Podcasts, YouTube, Amazon Music, and wherever you listen.

Audience: committing to the next 90 days of real, accountable, accessible practice is now your challenge. This is Human Signal. Thanks for listening. This is Dr. Floyd signing off.

================================================================================

END OF TRANSCRIPT

Slug: digital-accessibility-ai-era-dr-michele-williams

Published:: 2026

Blog post: humansignal.io/blog/digital-accessibility-ai-era-dr-michele-williams