Decisions issued: 17 September 2026
Scope: United Kingdom cases on Facebook
Reviewed: 21 September 2026
Two decisions from Meta’s Oversight Board show why identifying an AI-generated item is only part of deepfake literacy.
In one case, a realistic video falsely portrayed a Scottish local councillor making an inflammatory statement about refugees. In the other, an AI-generated video manipulated the likeness, words and actions of a young Muslim health campaigner as part of wider online harassment. The Board required Meta to remove both posts and recommended stronger treatment of harmful manipulated media, including more prominent warnings, reduced distribution and clearer protections for private individuals.
The Board is not a UK regulator, and its decisions do not create new duties for schools. They do, however, provide current UK examples of a practical problem: a platform label may be absent, disputed or insufficient to prevent harm. For educators, the cases support teaching pupils to examine provenance, context, intent and likely impact, and to know how to report harmful content – without replaying the abusive material itself.
Primary sources: UK politician decision and campaigner decision.