AI for Accessibility: Technology That Opens Doors

Some of the most meaningful AI uses are the ones that help people do things that were difficult or impossible before.

• Live captions and transcription for deaf and hard-of-hearing people, and for anyone in a noisy room.
• Image descriptions (alt text and scene descriptions) for blind and low-vision users.
• Text-to-speech and reading aids for dyslexia, low vision and fatigue.
• Voice control and dictation for people with limited mobility.
• Voice banking and cloning so people losing speech can keep a voice that sounds like theirs.
• Plain-language rewriting for cognitive accessibility and second-language readers.
• Real-time translation across spoken languages.
• Summaries and structure to help people who struggle with long documents.

• Add alt text to images (AI can draft it, you should check it).
• Provide captions for video and audio.
• Use clear headings and plain language.
• Check colour contrast.
• Don't rely on AI alone: test with real assistive technology and real users.

• AI descriptions can be wrong or miss important detail.
• Automatic captions need review for names and jargon.
• Involve disabled people in design and testing. They know what works.

Accessibility is where AI can have outsized benefit. Use it to widen access, and keep humans checking quality.