Watchdog Says Meta Needs to Do More to Protect Ordinary People from Sex-Based Deepfakes
Meta’s Oversight Board has urged the social media company to strengthen its protections against ordinary people who are targeted by sex victims. The board is recommending the addition of AI-generated impersonations to Meta’s Adult Sexual Exploitation policy, arguing that such images and videos are not automatically consenting. It also wants Meta to allow users to designate “linked accounts,” such as trusted friends and family, who can report serious violations such as inappropriate intimate photos on their behalf.
Finally, the Board recommends making AI-generated impersonation a separate category from harassment and nudity in company content reporting and complaint forms. Currently, only residents of Texas and Florida have access to a special form that lists fake depth photos as the reason for the report. The board wants all Meta users to have access to it, because “intimate abuse generated by AI without consent, including sexual impersonation, is a global problem.”
The Meta Oversight Board came up with these recommendations after an investigation into an incident where the company continued to ignore user reports about impersonating a friend for sex on Instagram. It launched an investigation after receiving a complaint from a user who reported an AI-generated video on Instagram showing a woman adjusting her dress, with her underwear visible in a few frames. According to the Board’s report, the journalist said that they were a friend of the person portrayed in the video without the content. The person featured in the AI video had already closed their account.
Two users initially reported the video to Meta, but the company did not remove the deepfake. The user who complained to the Board filed a complaint with Meta first, but the company did not remove the video from Instagram. After the Board itself raised the issue with Meta, the company simply made the post adult-only but concluded that it was not appropriate to remove it under its social standards.
Meta told the Board that when the post was originally reported, there was no indication that the person in the AI deepfake was a real person. If the person in the photo had reported the video, it would have violated its Elder Abuse Policy. Self-reporting would have served as a clear sign of disapproval. Other reliable indicators of disapproval in Meta’s eyes are reports from law enforcement, the media or trusted partners. Captions or page titles that suggest the images or videos are being shared in a “vindictive or extremist manner” will also work.
The board says that Meta’s responses to its investigation indicate that the only effective way for non-public members to find out that they disagree is to report them. It wouldn’t be easy for ordinary people to get law enforcement or the media involved, after all. Those ways are more accessible to people who are respected in the society. Meta is required to respond to these recommendations, but is not obligated to use them. If it chooses to accept them, the Board will monitor its performance. In the specific case that started this investigation, the Board overruled Meta’s decision to leave the video up and demanded that the company remove the post.
“It is clear that the scale, speed and complexity of AI tools have caused an increase in inappropriate sexual content produced by AI around the world. The spread of fake deep sex videos leads to reputational and psychological damage, which disproportionately affects women and girls, and has a negative impact on participation in social and political life,” wrote the Board in an investigative report.
This is not the first time the Board has criticized Meta on issues involving AI content and testing. Mid-2025, called the company’s inability to consistently enforce its rules “unconscionable and inexcusable.” Back in March, the Board urged Meta to create a new AI content rule separate from its misinformation policy. That recommendation came from an investigation involving an AI-generated video it claims shows damaged buildings in the Israeli city of Haifa. The video was posted by an account claiming to be a news outlet but was actually owned by a user from the Philippines.
