AI, Accessibility, and the Missing Sign Language: A Deaf Professional’s Perspective

Recently, Ofsted published two documents exploring AI in education. One outlines its approach to inspections: Ofsted won’t be inspecting AI tools directly, and they don’t expect AI to be used. But if AI is being used by staff or students, inspectors will examine how it’s managed and what impact it has, especially if there are risks to fairness, safeguarding, or data protection.

Their second document shares insights from early adopters: schools and colleges using AI for lesson planning, admin, SEND support. Others are struggling with misuse, staff concerns, and blurred ethical lines.

That’s important. But what about universities?

Who inspects how AI is used in higher education, for teaching, exams, support systems? I’ve heard from students and staff that many decisions are made at the top, without consultation, especially around tools like Zoom, MS Teams, or AI-based learning platforms. Often, accessibility is an afterthought.

Accessibility Must Include Sign Language

Last Friday, I was proud to share my session at Jisc:"Bridging Science: AI, BSL, and Inclusive Communication."

We discussed WCAG (Web Content Accessibility Guidelines) and the European Accessibility Act (EAA), taking effect in June 2025. But Sign Language is still overlooked in digital standards.

One excellent LinkedIn post reminded me:

“WCAG A, AA, AAA levels aren’t about ‘good, better, best.’ They’re about scope, not quality. True accessibility is meeting user needs, not chasing compliance.”

That’s exactly the issue. WCAG doesn’t require Sign Language. It’s seen as optional, even though Sign Language is a language. And when standards ignore language access, they ignore people.

Where AI Can Help—and Where It Shouldn’t

AI could improve Access to Work (ATW). I’ve had an ATW case since 2004, but each year I must re-explain everything. This year, my case manager asked for a video call just to ask what hearing aids I use. They're not audiologists. They don’t know my history. AI could store my record, track budgets, log tasks - things ATW staff often get wrong or delay.

But AI should never replace BSL interpreters. Interpreters understand Deaf culture, regional dialects, grammar, and nuances that no AI tool can replicate. BSL isn’t just a visual translation - it’s a rich, living language.

What We Need

• Ethical Standards for AI and Sign Language

• AI tools co-designed with Deaf professionals

• Deaf-led insights in policymaking (Ofsted, ATW, DfE, Jisc)

• Systems built with us not just for us

  
  

Too often, accessibility decisions are made by tech or leadership teams without input from the people most affected. That’s not inclusion, that’s convenience.

People first. Then technology.

🧭 We don’t need every Deaf organisation to be an AI expert. But we do need accountability, ethics, and proper representation, so AI serves us, not silences us.

Let’s keep pushing for meaningful, linguistic, human-centred accessibility.