AI oversight, bias & failure · Specialist
AI hallucination
Also called: artificial-intelligence hallucination
ELI5
An artificial-intelligence output that sounds confident but is unsupported, fabricated or incorrect.
Used in conversation
“The review flagged the AI hallucination in the way the people or material were represented.”
Pitch context
You may encounter “AI hallucination” in accessibility, localisation, ethics & ai conversations or materials.
Why it matters: Explains access, language adaptation, care, provenance and emerging synthetic-media responsibilities.
Sources & evidence · 8
Supporting source families. This entry is an editorial synthesis rather than a quotation from one canonical glossary.
- World Wide Web Consortium — Making Audio and Video Media AccessibleSupporting source family · primary · checked 2026-08-14
- Netflix — Partner Help Center Delivery SpecificationsSupporting source family · primary · checked 2026-08-14
- Coalition for Content Provenance and Authenticity — C2PA Technical Specification and Content CredentialsSupporting source family · primary · checked 2026-08-14
- SAG-AFTRA — Artificial Intelligence ResourcesSupporting source family · primary · checked 2026-08-14
- U.S. Copyright Office — Copyright and Artificial IntelligenceSupporting source family · primary · checked 2026-08-14
- Writers Guild of America West — Artificial Intelligence: Know Your RightsSupporting source family · primary · checked 2026-08-14
- Coalition for Content Provenance and Authenticity — C2PA Technical SpecificationSupporting source family · primary · checked 2026-08-14
- Coalition for Content Provenance and Authenticity — C2PA Conformance ProgramSupporting source family · primary · checked 2026-08-14