When technology learns to adapt to people
Accessibility often becomes visible in the simplest moments. When I enter a conference room or public space, I notice the podium height, walking path, seating, lighting and ergonomics — asking myself whether the space is accessible for my needs. Some needs are visible and permanent; others are temporary, situational or invisible. Accessibility is about dignity, independence and participation.
More than 1.3 billion people globally live with disabilities, and the need for accessibility is expected to grow as populations age. As education, work, healthcare, banking and public services become more digital, digital accessibility is no longer a niche need; it is a social and economic priority.
“The question is not whether AI will make technology more intelligent. The question is whether it will make technology more inclusive.”
At Microsoft, accessibility has been part of the journey for more than three decades. What excites me now is the shift AI enables: from people adapting to technology, to technology adapting to people.
AI that helps people see, hear, speak and act
Consider a blind professional working on a PC and receiving a presentation filled with charts, screenshots, images and complex spreadsheets. Earlier, understanding those visuals often depended on someone else or became a productivity barrier. With features like AI-powered image descriptions in Windows Narrator and experiences connected with Microsoft 365 Copilot, visual content can now become meaningful context — not just an image, but a description, an explanation and the start of informed action.
Speech is another powerful modality. Windows Voice Access enables people to control their PCs, open applications, navigate and create content hands-free. For someone with a motor disability, this can be life-changing. For many others, it points to a future where computing feels more natural, conversational and intuitive.
AI is also expanding communication. Imagine someone experiencing progressive speech loss who can preserve a version of their voice for future use. Azure AI Speech and Neural Voice technologies, including Microsoft’s work with partners such as Tobii Dynavox, can help people with speech disabilities communicate with more natural and personalised voices. This is about identity, individuality and self-expression.
“Build for one. Extend to many.”
This Microsoft Inclusive Design principle captures a truth we see often. Live captions designed for people who are deaf now help people in airports and open offices. Speech input, including Voice Access and Voice Typing in Windows, supports accessibility while helping others create and navigate more efficiently. AI-generated summaries support cognitive needs and help anyone facing information overload.
Designing for specific needs often leads to innovation that works for everyone.
Accessible content as a default
Perhaps the most scalable opportunity is helping everyone create accessible content by default. AI can flag missing alt text, recommend clearer structure, simplify language and support captions so documents, presentations and meetings become easier for assistive technologies to use. Accessibility becomes part of creation, not an afterthought.
Is AI itself inclusive? The real beginning
While AI can help reduce the digital divide, its impact depends on whether AI is designed to work for more people. Models are shaped by the data, feedback and design choices behind them. If disability experiences are underrepresented or misrepresented, AI systems may work less accurately for the communities they are meant to support.
This is why representative data, consented feedback and direct participation from disability communities are essential to building technology that works for everyone.
At Ability Summit, Microsoft shared its work and research collaborations focused on making AI work better for more people with disabilities. Work with Be My Eyes helps Microsoft learn from consented, privacy-conscious real-world experiences of blind and low-vision people. The Speech Accessibility Project, in partnership with the University of Illinois, is collecting consented, diverse speech samples from people with ALS, Parkinson’s disease, cerebral palsy and Down syndrome, with the goal of helping speech recognition better support non-standard speech patterns. This is just the beginning, and there is definitely more work across the industry to do, but the direction is clear.
Future: technology adapts to people
The future of AI could help people focus on what is most meaningful by reducing routine work and supporting productivity, creativity and personal choice in how they interact with technology and different modalities.
People with disabilities must not only be part of this journey; they must help co-create it. The principle is simple: Nothing about us without us, as we build a future where technology helps shrink the digital divide.
