I have lost count of the number of times a mainstream feature turned up years after blind people had already built their own version of it. Screen readers made computers speak long before “voice assistant” was a marketing term. The same pattern is repeating with AI, and once again blind AI pioneers are working out what these tools are actually good for.
We Are Blind AI Pioneers, Whether We Meant To Be Or Not
Necessity does that to a community. When built-in tools do not work for you, you either give up or build something that does, and blind and low vision people have spent decades doing the second thing. AI has turned out to be a genuinely useful new material to build with.
What strikes me is not that blind people are using ChatGPT, Claude, or Gemini, but that we are using them to close specific, longstanding accessibility gaps, often faster than the companies who owe us those fixes.
It is not the first time, either. Plenty of us were wearing Meta’s Ray-Ban smart glasses and asking their AI to describe a room, read a menu, or identify a product long before the glasses became a mainstream fashion item. We are always looking at the newest technology first and asking what it can do for us, rather than waiting for someone else to build the accessible version.
An Itch That Needed Scratching: Item Chooser for NVDA
Steven Scott, co-host of the Double Tap podcast, wanted something that NVDA users on Windows had envied in VoiceOver for years: a fast, searchable way to jump straight to any item in the current window or web page, much like VoiceOver’s Item Chooser (VO+I) on a Mac.
Rather than wait for someone else to write it, Steven used Claude as a coding partner, talking through the problem conversationally instead of writing code himself. The result is Item Chooser for NVDA, free on GitHub and activated with NVDA+I.
Steven is not a professional developer. He identified a gap, described what he needed in plain language, and ended up with a working NVDA add-on that other blind users now rely on daily.
Quill: A Text Editor Built By Someone Who Actually Uses a Screen Reader
Jeff Bishop has taken a bigger swing with Quill, a free, open-source writing environment for Windows and macOS built with screen reader use as the starting design requirement, not an afterthought.
Every command in Quill announces its outcome, and nothing depends on visual feedback. It bridges the gap between Notepad and a full code editor, and it includes an AI writing assistant with a review-first model, so nothing changes in your document without your approval.
Jeff talked through Quill’s design on Double Tap, explaining how it “meets users where they are” through configurable profiles, from a clean beginner view to a full developer toolkit. A good example of AI used inside a tool, not a tool generated by AI.
ClipMan: The Clipboard Tool Sighted Users Never Think About
Andre Louis, a blind musician, created ClipMan, a free clipboard manager for Windows and Mac that keeps a searchable history of everything copied.
It sounds small until you need it. Sighted users rarely think about their clipboard because they can see what they last copied. ClipMan gives that same casual confidence to screen reader users, with pinned entries, quick-paste hotkeys, and syncing across machines.
ClipMan came up alongside Quill on Double Tap, in the same episode where accessible gaming developer Robin Christopherson dropped in to praise it. A good reminder that genuinely useful accessible software often comes from individuals solving their own problems first.
Learning AI Itself, From Someone Who Lives It
None of this helps if blind users do not know how to operate ChatGPT, Claude, or Gemini with NVDA, JAWS, VoiceOver, or TalkBack. That is the gap Tony Gebhard, a totally blind musician and assistive technology professional, has written a book to fill.
No Sight Required: The Blind User’s Guide to AI is a practical, no-gatekeeping walkthrough written by someone who uses these tools every day without sight, covering every major AI assistant across Windows, Mac, iPhone, and Android with screen reader commands spelled out in plain words. It is available on Amazon in ebook and paperback (Affiliate Link), and Tony sells accessible ePub, MP3, and Word editions directly if Kindle does not suit you.
Why Blind AI Pioneers Keep Turning Up First
I do not think this is a coincidence. Blind AI pioneers keep turning up early in every new wave of technology because we cannot afford to wait for mainstream accessibility, and AI lowers the barrier that has always stood between a good idea and a working tool.
Steven did not need years of software training to build Item Chooser. Jeff did not need a large team to build Quill’s screen reader parity. None of it would have looked the same three years ago.
The same community keeps shipping more: Kelly Ford’s QuickMail, a screen-reader-friendly email client for IMAP servers, and Claude Code Reader, an accessible wrapper around Anthropic’s Claude Code with native VoiceOver and NVDA/JAWS support. Both point to where this is heading: accessible interfaces built around AI tools, not just tools built with AI.
What Blind AI Pioneers Mean for the Rest of Tech
The pattern goes well beyond screen readers and text editors. Configuration software for things like Keychron keyboards or Elgato Stream Decks is often visual-only, with no keyboard access and no screen reader labels, leaving blind users to guess or ask a sighted friend for help. Community members are now using AI to reverse-engineer that software and build their own accessible configuration tools, closing gaps the manufacturers left open.
Which raises an obvious question. If a blind user with no engineering background can fix a barrier like that with AI in an afternoon, why did the company selling the product not build it accessibly in the first place? The wider tech industry still treats accessibility as a compliance checkbox far too often, rather than a source of genuine innovation, and AI is quietly exposing how thin that excuse has always been.
If you are blind or low vision and have not tried an AI assistant on a specific accessibility gap in your own workflow, that is where I would start. You do not need to be a developer, just a problem worth solving and the patience to describe it clearly.
Tell Me What You Think
Have you built something with AI to solve your own accessibility problem, or found a community tool that has genuinely changed how you work?
Tell me what you think in the comments below or on X @timdixon82
Sources
- No Sight Required by Tony Gebhard
- Buy No Sight Required on Amazon (Affiliate Link)
- Item Chooser for NVDA by Steven Scott
- Quill by Jeff Bishop
- Why Quill Is Changing Text Editing for Blind Users – Double Tap
- ClipMan by Andre Louis
- Quill, Clipman And New Book: No Sight Required To Help You Learn AI – Double Tap
- Claude Code Reader
- QuickMail by Kelly Ford
