Meta’s Oversight Board to Decide on Regulation of AI-Generated Sexual Images

Meta’s Oversight Board is tackling the company's policies regarding AI-generated content, specifically focusing on explicit images featuring public figures. The board has accepted two notable cases involving AI-created explicit imagery.

Although Meta’s existing guidelines prohibit nudity on Facebook and Instagram, the board aims to evaluate the effectiveness of these rules and their enforcement against explicit AI-generated imagery. Often termed “deepfake porn,” this content has increasingly emerged as a form of online harassment, prompting a surge in regulatory proposals. Through these cases, the Oversight Board may encourage Meta to establish enhanced measures to combat such harassment on its platforms.

The board has chosen not to disclose the identities of the two public figures involved in order to prevent further harassment. However, it provided context about each situation. The first case concerns an Instagram post depicting an AI-generated nude image of an Indian woman, shared by an account dedicated solely to AI-generated images of Indian women. Although the post was reported to Meta, the report was closed without review after 48 hours. The user’s appeal also went unreviewed. Ultimately, Meta removed the post following an appeal to the Oversight Board, which agreed to handle the case.

The second case involves a Facebook post within an AI art group, showcasing an AI-generated picture of a nude woman with a man groping her. This image was intended to resemble an American public figure, whose name was included in the caption. The post was automatically taken down due to prior reports, and despite the user appealing the removal, the appeal was closed without consideration. The user subsequently appealed to the Oversight Board, which agreed to review this case.

Co-chair Helle Thorning-Schmidt emphasized the board's focus on evaluating potential disparities in how Meta applies its policies across different regions. "We know that Meta is quicker and more effective at moderating content in some markets and languages than others,” she remarked. “By examining cases from both the US and India, we aim to determine whether Meta ensures fair protection for all women globally.”

The Oversight Board is soliciting public feedback over the next two weeks and plans to announce its decision and policy recommendations for Meta shortly thereafter. A recent similar case involving a misleadingly-edited video of Joe Biden resulted in Meta committing to better labeling of AI-generated content on its platform.