The independent Oversight Board for Meta is calling for a major shift in how the social media giant handles non-consensual sexual imagery. In a new report, the board argues that Meta’s current policies fail to protect everyday users from the growing threat of AI-generated sexualized deepfakes. While the platform has strict rules against revenge porn, the board believes those guidelines fall short when applied to the hyper-realistic and malicious content now flooding the internet.
Currently, Meta’s policies generally require users to report non-consensual sexual content before the company takes action. The Oversight Board claims this approach is entirely ineffective for victims of deepfakes, who often do not realize their images have been manipulated until the content has already spread to thousands of people. The board urges Meta to move toward a more proactive system that identifies and removes these harmful images before they cause irreparable damage to a person’s reputation or mental health.
The board’s recommendations follow a specific review of cases involving manipulated images that bypassed existing automated detection systems. Members of the board highlighted that bad actors now use readily available, low-cost AI tools to create high-quality sexualized content in mere minutes. Because these tools have become so accessible, the sheer volume of fake imagery has skyrocketed, leaving individual users to fight a losing battle against viral, non-consensual content.
To address this, the Oversight Board wants Meta to overhaul its “Non-Consensual Sexual Imagery” policy. They are pushing for the company to treat manipulated deepfakes with the same level of urgency as real imagery. The board suggests that Meta implement advanced scanning technologies that specifically target the biometric signatures of non-consensual deepfakes. By catching this content at the point of upload, Meta could potentially prevent millions of views before the material gains traction.
Beyond just technical fixes, the board believes Meta needs to be more transparent about how it handles these reports. They recommended that the company publish data on how often it detects sexualized deepfakes versus how often users report them. This level of accountability would show exactly how much of the problem remains hidden. The board noted that even a 5% improvement in detection accuracy could significantly lower the trauma inflicted on innocent users who find themselves targeted by anonymous attackers.
The economic reality of running a global platform complicates these efforts, but the board emphasizes that safety cannot take a backseat to user engagement. While it might cost tens of millions of dollars to build and maintain an effective AI-moderation system, the board argues that the current lack of protection creates a toxic environment that drives users away. They expect Meta to provide a formal response to these recommendations within the next 60 days, outlining a concrete plan for policy updates.
Ultimately, this push from the Oversight Board highlights a critical turning point for social media regulation. As generative AI makes it easier than ever to violate personal privacy, platforms like Facebook and Instagram face mounting pressure to protect their users from digital harm. Whether Meta will embrace these strict, proactive changes or continue to rely on a reactive reporting system remains to be seen. For now, the message from the board is clear: the era of waiting for victims to report their own abuse must end.
