Meta's Oversight Board will examine deepfake video of a UK politician Facebook left online
Published: (May 21, 2026 at 11:31 AM EDT)
3 min read
Source: Engadget
Source: Engadget
PJ McDonnell/Shutterstock
Overview
Meta’s Oversight Board is investigating an apparently AI‑generated video of a British Labour Party councillor from Scotland that misrepresents the politician’s views on immigration. The video, posted as part of an album on Facebook in November 2023, shows the councillor making sarcastic and offensive comments about refugees and sexual assault.
The post also includes:
- A second video of pro‑Palestine protesters that the Board suspects is AI‑generated.
- A likely genuine still image of several women, including the councillor, holding anti‑far‑right signs.
- A caption that makes baseless accusations of tax evasion against the politician.
Board Findings
- Engagement: The post received relatively minimal interaction and contained no AI labels or disclosures.
- User Reports: Two users claimed the content violated Meta’s Bullying and Harassment policy. Meta’s systems did not forward the complaints for human review, so the content remained online. One of the reporters appealed to the Oversight Board.
- Context: The councillor has been vocal against racist misinformation used to fuel anti‑migrant sentiment during protests over housing asylum seekers in UK hotels. They have previously reported threats and AI‑generated defamation.
Meta’s Position
- Public‑Figure Exception: Meta said the post was not removed because the politician is an adult public figure, which does not automatically trigger protection from “unwanted manipulated imagery.”
- Policy Application:
- The Board’s claim that the video violated the Hateful Conduct policy was rejected; Meta argued the content did not meet the criteria for misinformation and was not flagged by any member of its Trusted Partner program.
- The content was not deemed election interference, as the councillor did not stand in Scotland’s recent local elections.
- Low engagement and the “satirical” nature of the video were cited as reasons for not applying an AI label.
Public Comment Period
The Oversight Board is inviting public comments on the case, which may inform future policy recommendations to Meta. While the recommendations are not binding, Meta must respond within 60 days. The Board operates as an independent external body, though it relies financially on Meta.
Related Cases
- Irish Presidential Candidate Deepfake (2023): An AI‑generated video of Catherine Connolly, then a presidential candidate, falsely announced her withdrawal from the race. The video was shared nearly 30,000 times before Meta removed it after being alerted by the Irish Independent. Connolly called the video “a disgraceful attempt to mislead voters and undermine Ireland’s democracy.”
- Joe Biden Deepfake (2022): Meta previously faced criticism for allowing a maliciously edited video of President Biden to remain on the platform, highlighting ongoing challenges with politically motivated deepfakes.