New York City Bar + TestParty: AI and Accessibility

AI and Accessibility: Reflections on My Conversation with the NYC Bar

On June 13, 2025, the New York City Bar Association released a podcast episode titled “AI and Accessibility.” I’m proud to share that I was featured in this discussion as part of my role on the Presidential Task Force on Artificial Intelligence and Digital Technology, specifically the Subcommittee on the Impact of Artificial Intelligence on People with Disabilities and Underserved Communities.

Because the episode touches directly on the intersection of AI, law, and disability inclusion—the same intersection where TestParty operates every day—I wanted to highlight it here for our community. It’s an important conversation, and one I’m honored to be a part of.

The podcast builds on a new Task Force report examining how AI systems currently affect people with disabilities, where those harms are already emerging, and what policymakers, technologists, and advocates need to do next.

Listen to the Podcast

If you'd like to hear the full discussion, you can listen on Apple, Spotify, YouTube Music, and iHeart

A new City Bar report from the Presidential Task Force on Artificial Intelligence and Digital Technology highlights the harmful effects of artificial intelligence (AI) on people with disabilities, and the harm likely to occur for them in the future. The Task Force’s  Subcommittee on the Impact of Artificial Intelligence on People with Disabilities and Underserved Communities continues the conversation in this podcast episode featuring attorneys, researchers, and technology leaders. They talk about the current challenges and opportunities AI presents for people with disabilities, emphasizing the importance of inclusive design and representation.

Why This Conversation Matters

I’ve said it before: AI is shaping the structure of society faster than regulations, institutions, and social norms can keep pace. For disabled people—17% of the world—AI can be a powerful equalizer or a silent barrier.

In the podcast, I joined attorneys, researchers, and accessibility leaders to break down both sides of that reality:

• Where AI is quietly disadvantaging disabled people
• Where AI can meaningfully expand access
• And the real work required to build systems that don’t leave people behind

For anyone working in accessibility or AI (or, frankly, any company deploying automated systems), this is the kind of conversation that reveals blind spots and inspires better design.

How This Fits into the City Bar’s Broader AI Work

The “AI and Accessibility” episode isn’t a one-off. It’s part of a larger, ongoing effort by the New York City Bar’s Presidential Task Force on Artificial Intelligence and Digital Technologies to:

• Analyze the social and legal impacts of AI
• Publish reports and guidance
• Host CLEs, institutes, and webinars on emerging AI issues
• Create space for cross-disciplinary collaboration

That broader context matters: AI and disability inclusion aren’t niche topics—they’re increasingly central to how courts, regulators, and practitioners think about fairness, equity, and the rule of law in a digital age.

Who’s Around the Table?

The episode brings together a cross-section of people who don’t often get to talk in the same room:

• Lawyers and policy leaders thinking about the ADA, emerging AI regulations, and how courts will handle new kinds of harms.
• Researchers and technologists who study AI systems and build the tools that are rolling out into products.
• Disability and accessibility advocates who keep the conversation anchored in lived experience, not just abstract “users.”

That mix matters. When AI tools touch everything from hiring and housing to healthcare and education, getting law, technology, and disability communities aligned is no longer a nice-to-have—it’s the only way to design systems that deserve trust.

What We Covered (Timestamps Included)

The podcast is structured around several deep-dive segments:

The podcast is organized into natural “chapters” that you can follow using the timestamps in the episode description.

1. Overview of AI’s Impact on Disabilities (03:06)

The conversation starts with the big picture: AI is already classifying, ranking, and filtering people in ways that can either open doors or quietly close them.

The group talks about:

• How algorithmic decision-making shows up in hiring, benefits, education, and more
• Why disabled people are often “invisible” in the datasets that power AI
• The risk that existing bias gets encoded and scaled, rather than corrected

It sets the tone: this isn’t a hypothetical future problem. It’s happening now.

2. Building a Disability-Inclusive AI Ecosystem (04:44)

From there, the speakers zoom out and ask: what would a disability-inclusive AI ecosystem actually look like?

A few themes emerge:

• Disabled people are not edge cases; they’re core stakeholders.
• Inclusion has to happen across the lifecycle: research, design, development, evaluation, and governance.
• Representation matters—both in the data and in the teams building and deploying AI.

This is the “north star” section of the episode, and it’s very aligned with how we think about product-building at TestParty: accessibility can’t be bolted on at the end.

3. Generative AI and Representational Harms (09:08)

Next, the podcast dives into generative AI—the models that write text, generate images, and increasingly shape how disability is portrayed online.

The conversation covers:

• How models can reinforce stereotypes (“inspiration porn”, medicalized framing, or pity narratives).
• The ways disabled identities are erased or oversimplified in training data.
• Why representation isn’t just about imagery; it affects how systems respond to and support real people.

For anyone working with text or image models, this portion is a powerful reminder: the stories our systems tell about disabled people have real-world consequences.

4. AI in Digital Accessibility: TestParty’s Mission (15:54)

One segment of the podcast focuses on AI in digital accessibility, including the work we’re doing at TestParty.

Key ideas discussed:

• Why so much of the internet still fails basic WCAG guidelines—even in 2025.
• The role of AI in scaling accessibility work, especially for tasks like:
  • Generating meaningful alt text
  • Remediating PDFs
  • Flagging and fixing common code-level issues
• The importance of human-in-the-loop design: AI to accelerate and support experts, not replace them.

This part of the conversation highlights a core belief behind TestParty: AI is most powerful when it’s pointed at concrete, fixable problems in the code, with humans defining what “good” actually looks like.

5. Challenges and Legal Perspectives (19:39)

The lawyers in the group then unpack the legal side:

• How existing laws (like the ADA) interact with AI-driven systems
• Where current frameworks fall short when harms are algorithmic or opaque
• The tension between encouraging innovation and enforcing accountability

The takeaway: law is moving slower than AI, but it’s moving—and organizations building or buying AI tools need to keep a close eye on how those obligations evolve.

6. Inclusive Design and Evaluation in AI (35:26)

Later in the episode, the discussion shifts to design and evaluation:

• “Nothing about us without us” as a principle for AI projects
• Why accessibility testing must include disabled users and assistive technologies
• The need for better benchmarks and evaluation methods that explicitly consider disability

If you’re a product manager, designer, or developer, this portion is practically a playbook for how to weave accessibility into your AI work in a serious way.

7. Final Thoughts and Future Directions (44:54)

The episode closes with forward-looking reflections:

• What gives the panelists hope about AI and disability inclusion
• Where they see the biggest risks if we “set and forget” these systems
• The kinds of collaboration they want to see between industry, advocates, and policymakers

The message is clear: AI can be part of the problem or part of the solution. The difference is whether disabled people are centered from the beginning.

My Key Takeaways

Thinking back on this conversation, a few insights stand out:

• AI cannot be neutral—it reflects the data and decisions behind it.
• Disability is not an afterthought; it must be a first-order consideration in AI design.
• Automation is most powerful when paired with human expertise, especially in accessibility work.
• Policy conversations are finally catching up, but we still have a long way to go.
• And most importantly: inclusive AI isn’t a technical challenge—it’s a leadership challenge.

These principles shape the work we’re doing at TestParty every day.

I’m grateful to the NYC Bar for elevating this topic and thrilled that accessibility is becoming a central part of the national AI conversation. We’ll continue pushing forward—building tools, shaping policy, and ensuring that the future of AI is one everyone can participate in.