Shipping Accessibility with AI: If AI writes the code, do we still need accessibility testing?

Deque Systems on YouTubeWatch on YouTube

SaveMark complete

Sign in to save videos and track what you’ve finished.

IntermediateAccessibility25 min

About this video

The video argues that AI coding agents do not produce accessible code by default, and that the testing setup around the model strongly affects output quality. It covers decomposing WCAG into small steps and the continued need for human review of complex edge cases. It suits developers and accessibility practitioners who already work with WCAG.

What you’ll learn

  • Explain why AI-generated code still needs accessibility testing
  • Break WCAG requirements into small steps for a coding agent
  • Structure a closed feedback loop for AI accessibility checks