The Oversight Board, often described as Meta’s "supreme court", has instructed the company to take down two AI-generated videos from Facebook and to bolster protections against manipulated imagery. In its ruling, the board criticised what it called 'inadequate' safeguards that allowed the deepfakes to remain on the platform.
One of the contested clips showed a Labour councillor in Scotland apparently making inflammatory remarks about refugees; the board said that video should not have been left accessible on Facebook. A second AI-generated clip depicted a young Muslim campaign volunteer giving health advice while engaging in bizarre exercises or consuming junk food — an image the board found to be a false portrayal that also should have been removed.
The decision underscores growing concern about the ease with which synthetic media can be used to misrepresent public figures and private citizens alike. The board's finding comes amid broader calls from politicians, campaigners and civil society groups for tighter controls on AI-driven disinformation and clearer rules from major tech platforms.
While the board ordered removals in these specific instances, it also emphasised the need for Meta to strengthen its systems to detect and respond to AI-generated content more proactively. The ruling points to potential changes in how platforms balance free expression with the risks posed by synthetic media in political and social contexts.
Meta has in the past said it relies on a mix of automated detection and human review to police content, and the Oversight Board's critique highlights continuing gaps as generative AI tools become more sophisticated. The ruling is likely to fuel further debate about platform responsibility, technology safeguards and what regulatory or industry measures are needed to curb the spread of convincing fakes online.