Write inclusive, people-first language about AI
summary
A proposal called "people first accessibility language for AI and agents" argues that people-first language must center humans because accessibility is for people not machines; LLMs and AI agents (model, rules, memory, tools) can help but cannot guarantee accessible outputs. Key numbers: 16% of the world has a disability and 95.9% of homepages fail automated accessibility checks; a 1,600-person survey reports 37% EU Act compliance, 75% link accessibility to revenue, and 86% expect AI to solve it. Chrome’s Modern Web Guidance is explicitly harm reduction, not a guarantee.
AI-written summary · about
chapters (8)
- 0:00Inclusive AI language intro
- 3:12Accessibility and AI defined
- 11:44AI cannot solve accessibility
- 16:57Anthropomorphism in AI terms
- 24:23Human-centered language
- 26:49People-first AI language
- 28:13AI tools documentation
- 31:05Q&A on AI terms
related talks
52mPreparing Accessibility for the Future and the Present
48mAI and Disability: Progress or Exclusion?
41mProgramming for Accessibility
50mHuman-Centered AI and Accessibility
49mUsing AI to Streamline Inclusive Design Without Losing Humanity
28mBuilding in accessibility for broader value to all
34mAI for Accessibility: Empowering Inclusive Digital Experiences
55mBridging the Digital Divide: Empowering Accessibility Through AI Innovations