Meta: Oversight board demands clearer rules for sexual deepfakes

Key Points

1. Call for Stricter Guidelines: Meta’s Oversight Board urges precise rules for handling sexual deepfakes.
2. Impact on Women: Sexual deepfakes disproportionately harm women and girls.
3. Reporting Issues: The current reporting system needs improvement for quicker and more efficient responses.

Meta’s Oversight Board, an independent body tasked with regulating the platform, is calling for clearer and stricter rules regarding sexual deepfakes. Composed of members from research and non-profit organizations, the Board emphasizes the need for Meta to address the rising issue of AI-generated sexual content more effectively.

Cases Highlighting the Issue

In recent instances, AI-generated images of women from public life in India and the United States were shared without consent. These sexually explicit deepfakes are particularly harmful to women and girls. The Board stresses that Meta must take stronger actions to regulate such content and make it easier for users to report and remove these images.

In one case, a woman from India discovered a fake nude image of herself circulating on Instagram. She reported it as pornographic, but the report was automatically closed after 48 hours due to processing delays. An appeal she filed later was also closed automatically. Frustrated, she took her complaint to the Oversight Board, which ruled that Meta should have removed the image for violating community guidelines.

Recommendations for Meta

The Oversight Board has made several recommendations to Meta:

1. Update Guidelines: The Board suggests that Meta needs to update its guidelines to reflect the capabilities of current AI technology. The current prohibition against derogatorily edited images using Photoshop is outdated. Meta should explicitly ban the creation or editing of sexual content without consent.
2. Reclassify Content: The Board believes that such cases should fall under the “Adult Sexual Exploitation Community Standards” rather than being classified as “Bullying and Harassment.” This reclassification acknowledges that these images are often used to lure users to external sites and do not solely serve the purpose of bullying.
3. Rename Category: They recommend renaming the category to “Non-Consensual Sexual Content” to better reflect the nature of the issue.
4. Improve Reporting Mechanism: The automatic closure of reports after 48 hours is problematic. The Board insists on a more robust system that ensures such reports are processed and addressed promptly.

Challenges Ahead

Generative AI is making it easier to create realistic deepfakes, presenting new challenges for platforms like Meta. The Oversight Board’s recommendations aim to improve the platform’s ability to handle these issues effectively. However, the rapid advancement of AI technology means that continuous updates to policies and guidelines will be necessary to keep pace with new developments.

Meta’s Oversight Board is pushing for these changes to protect users, particularly women and girls, from the harm caused by non-consensual sexual content. By implementing these recommendations, Meta can create a safer online environment and ensure that their policies are equipped to handle the complexities introduced by AI-generated media.

//Source//