The Oversight Board, a global body of experts that reviews Meta’s most difficult and significant decisions related to content on Facebook, Instagram, and Threads, published its findings regarding the topic of the explicit, non-consensual, AI-generated images of female public figures shared on Facebook.
The Board analyzed two similar cases that were originally treated differently by Facebook (though they are relevant for Instagram and Threads, too). Both cases involved public female figures, one in the U.S., and one in India. The expert panel found both instances of non-consensual deepfakes, unsurprisingly, to violate Community Standards; however, the Indian case was originally not deleted by Facebook. Instead, the user report was automatically dropped after 48 hours. This was a hint that Meta is not treating all global markets equally – one of many reasons why the company is criticized by experts and journalists.
Despite being an independent entity, the Oversight Board has significant authority that earned it the nickname “Meta’s supreme court.” It can overturn the company’s decisions in particular cases and issue policy recommendations to Meta.
In this case, the Board recommends that Meta:
- Move the prohibition on “derogatory sexualized photoshop” into the Adult Sexual Exploitation Community Standard.
- Change the word “derogatory” in the prohibition on “derogatory sexualized photoshop” to “non-consensual.”
- Replace the word “photoshop” in the prohibition on “derogatory sexualized photoshop” with a more generalized term for manipulated media.
- Harmonize its policies on non-consensual content by adding a new signal for lack of consent in the Adult Sexual Exploitation policy: context that content is AI-generated or manipulated. For content with this specific context, the policy should also specify that it need not be “non-commercial or produced in a private setting” to be violating.
Apart from these recommendations, the Board is concerned about the auto-closing of appeals for image-based sexual abuse, saying: “Even waiting 48 hours for a review can be harmful given the damage caused.” The Board considers this an issue that could have a significant human rights impact, requiring risk assessment and mitigation.
The analyzed cases uncovered also another problem regarding the so-called Media Matching Service Bank, which gathers previously deleted content and helps to automatically remove rules-violating images upon their republishing.
Meta explained that it relies on media reports, for example, an article informing about leaked images of a specific celebrity. “This can be useful when posts involve public figures but is not helpful for private individuals. Therefore, Meta should not be over-reliant on this signal,” says the Board.
These recommendations are not binding, although Meta must officially respond within 60 days. It often leads to changes in Meta’s internal policies.