AI tools offer new accessibility options but require cautious adoption, expert warns
Generative AI is helping people with disabilities interact with information and their surroundings, but Léonie Watson of TetraLogical cautions that these tools are no substitute for inclusive design and must be developed with user involvement.
This article was drafted with AI assistance from multiple sources and was reviewed and approved by a human editor before publication.
Generative AI is beginning to address long-standing accessibility challenges for people with disabilities, according to Léonie Watson, Co-Founder of TetraLogical, writing in a recent analysis. Watson notes that roughly 1.3 billion individuals globally have some form of disability, with around 16 million in the UK—about a quarter of the population. Despite progress in inclusive design, many digital services remain hard to use for those relying on screen readers or requiring simplified content, she says.
Watson describes how AI can help neurodivergent users by condensing lengthy documents into summaries, rewriting content in plain language, and answering clarifying questions. This turns reading into an interactive process where users control the pace and tailor information to their needs. For blind and low-vision users, AI can extract text from inaccessible PDFs and generate descriptions for images, diagrams, or everyday items like restaurant menus that lack alt text.
Real-time visual description is another growing application. AI integrated into smart glasses or mobile apps can analyze live video to identify objects, read signs, or describe a user's surroundings, potentially increasing independence. Watson acknowledges hurdles such as battery life, privacy, and consistency that still need resolution.
Watson emphasizes that AI outputs should not be trusted without scrutiny. She points out that image descriptions may include hallucinations—confident but incorrect details—and document summaries can contain factual errors or fabricated content. Users should treat AI as an assistant rather than an authority, she advises.
Recent UK research cited by Watson reveals that over one-third of disabled people worry about being left behind as AI becomes more prevalent in healthcare and public services. Additionally, many fear that without the involvement of disabled individuals in the design process, these tools will not meet their needs.
"When the dust settles on the hype around AI, it will not be just about what this technology can do, but also about who it is built with, and how it is made available," Watson writes. She concludes that inclusion requires deliberate planning, listening, and a readiness to challenge assumptions.