Facebook Oversight Committee Wants Better Rules Against Sexual Fraud

Facebook’s Oversight Committee has published a memo calling for changes to the social network. The independent group is specifically calling for updated text regarding deepfakes generated in sexual contexts.

“Intimate deepfake images disproportionately affect women and girls — undermining the right to privacy and to protection from physical and mental harm. Restrictions on this content are legitimate to protect individuals from the creation and dissemination of non-consensual sexual images, the Committee’s text says.

The topic turned into a debate after two incidents resulting in the publication of manipulated images By artificial intelligence (AI) involving women from the United States and India. In addition, other cases have emerged on these and other platforms featuring montages of women without clothes or with adult content.

In the first report, an AI montage featuring a prominent figure in India went viral on Facebook. Despite complaints, the post remained up for over 48 hours. Meta removed the posts only after a “reprimand” from the committeeHowever, no problems with the images were initially noted.

In the second case, involving a North American user, the image automatically removed because the assembly was already included in a manipulation image bank predefined.

What should change on Facebook?

In general, the group is Meta Replace the word “degrading” with “non-consensual” in the rules. Text Also, remove the word “Photoshop” and replace it with more general terms. Containing doctored media, not to mention Adobe software.

Also the platform New content needs to be added to strengthen the ban “Note that with AI-generated content, it doesn’t have to be just the content that’s infringing.”produced in non-commercial or private environments“.

According to the committee, the seriousness of the issue makes the removal of the content “a significant matter” the only effective way to protect those affected“. Facebook’s current tool on the subject, labeling images generated using artificial intelligence, is seen as inappropriate and ineffective because it keeps posts afloat.

The committee has been around since 2021 and has already criticized the company on several occasions, including for its lack of transparency. To date, Meta (which owns Facebook) has not officially commented on the case.