AI & disability

If your AI systems aren't operable and interpretable for people with disability, you're building exclusion into the system.

AI & disability

Once AI mediates services, disability inclusion becomes a systems question. If the interface is unusable, or the model misunderstands context, the outcome is exclusion, even when nobody intended it.

What this stream covers

  • Where AI systems commonly exclude people with disability
  • Assistive tech failure modes (chat UIs, forms, auth flows, “smart” help)
  • Decision systems: bias, interpretability gaps, and no meaningful review path
  • Governance requirements that prevent “oops” harm

If your organisation says “inclusive by default”

Then at the very least, your AI systems needs:

  • Governance records that back up that claim.
  • Test evidence that demonstrates operability for people with disability.
  • Remediation logs when it failed, and how you fixed it.
  • Accessibility acceptance criteria
  • Tested pathways