Continuous AI for Accessibility: Build Better, Faster, Inclusive Software with GitHub
summary
Open-source GitHub Actions pipeline runs axe-core scans, files GitHub issues, and can assign Copilot to open PRs that apply fixes (example: automated PR added an H1; one run reports six findings). Scanner supports auth, workflow-dispatch or PR triggers, and outputs cacheable results; it lives on the Marketplace and repo. Demos show flaky live runs and noisy mass-PRs if run across thousands of URLs, so teams use manual triggers, Copilot PR fixes, and custom instruction prompts to control scale.
AI-written summary · about
chapters (16)
- 0:00Introduction and welcome
- 1:05Speaker introduction and background
- 2:09Why AI matters for accessibility
- 3:31GitHub’s accessibility approach
- 5:55Continuous AI for accessibility explained
- 6:37Find file fix workflow demo
- 9:24Accessibility scanner overview
- 12:55Live demo setup and execution
- 21:20Demo challenges and troubleshooting
- 30:13Demo video playback and explanation
- 34:36Key takeaways and future plans
- 35:59Open source accessibility summit
- 37:04Q&A on testing scope and workflows
- 41:20AI’s role in accessibility testing
- 45:11Workflow automation and human oversight
- 48:18Closing remarks and resources
