Strengthen Bullying And Harassment Policies Against AI-Imagery Of People Not In Public Eye
17 सितम्बर 2026
The Oversight Board is requiring that Meta remove from Facebook an AI-generated video that bullies a young Muslim woman by manipulating her likeness and was part of a viral series of similar posts across social media. The Board finds that Meta’s bullying and harassment policies are inadequate to address AI-manipulated imagery of people not normally in the public eye.
Why This Matters
This decision comes as generative AI creates new risks for women and girls, with abuse increasing in scale, speed and severity. According to a 2025 report by UN Women, the lead United Nations (UN) entity on gender equality, many attacks are “often deliberate and coordinated and aimed to silence women in public life.” AI-generated videos are changing the nature of online harassment and have significant impacts on the rights of Meta’s users, including the public participation rights of Muslim women, and women and girls generally. These attacks often falsify not just an individual’s appearance but also their actions and words in a manner that is damaging to them.
About the Case
In 2025, a young Muslim woman in Europe was interviewed on TV as a campaign volunteer for improving menstrual health education and reducing stigma around health conditions for women and girls from ethnic minority backgrounds. In the following months, accounts across social media platforms digitally manipulated the interview footage to ridicule her by creating an apparent disparity between her role promoting health and her appearance, in an obvious attempt to suggest that she is an unfit messenger on health matters. These videos use her likeness to manipulate her actions and words. The AI-generated videos and images together garnered tens of millions of views across platforms, including Meta’s.
The Board selected a Facebook post, submitted by user appeal, as just one example of the abuse that unfolded for this woman. The post includes an AI-generated video of the woman introducing herself as her country’s new health ambassador. This false claim is replicated in the caption, alongside a laughter emoji. Her likeness is shown saying that she is going to demonstrate how to stay in shape and then is depicted engaging in a range of limited or absurd exercises and giving advice about healthy eating while consuming junk food. Multiple comments beneath the video from various users ridicule the woman, mainly for her weight and for wearing a hijab.
For the user who appealed to the Board, both their initial report to Meta and subsequent appeal to the company were automatically closed without being prioritized for review. Meta has since confirmed to the Board that its decision to ignore the post, leaving it on the platform without attaching an AI label, was consistent with its policies. Only when the Board asked questions about four disparaging comments on the post did Meta take action to remove the comments, confirming they violated the Bullying and Harassment prohibition on dehumanizing comparisons.
Key Findings
A majority of the Board finds that the video should be removed because it violates Meta’s Bullying and Harassment policy. Specifically, it contains a statement of inferiority about a private individual’s physical appearance, used AI to falsify her words and actions without her consent, and was part of a campaign that was experienced by the target as mass harassment.
For the majority, Meta’s definition of “unwanted manipulated imagery,” which is one of the bullying and harassment prohibitions, is too narrow, and should include falsifying a person’s actions and words, in addition to altering a person’s appearance.
Meta also defines mass harassment campaigns too narrowly, according to the majority, and the company should adopt new measures to reduce the burden on victims reporting them. These measures include adding violating unwanted manipulated content to the company’s Media Matching Bank so identical and near-identical content is removed, and using signals of AI-generated or manipulated media as an indicator of severity to prioritize reports for review.
A minority of the Board found the content did not meet Meta's definition of Bullying and Harassment, and that leaving the content up conformed with the company's human rights responsibilities.
The Oversight Board’s Decision
The Board overturns Meta's decision to leave up the content, requiring the post to be removed.
The Board stresses the importance of its previous recommendations calling for allowing third-party accounts to report bullying and harassment content.
Additionally, it recommends that Meta:
- Broaden the public definition of “unwanted manipulated imagery” to include deepfakes of a private individual saying or doing things they did not say or do.
- Add instances of violating “unwanted manipulated imagery” to Media Matching Banks for removal.
- Use signals of AI-generated or manipulated media as an indicator of severity for the purpose of prioritizing reports for review.
Futher Information
To read public comments for this case, click here.
