Meta Faces Renewed Pressure From Oversight Board Over AI-Generated Content

The Oversight Board has renewed its criticism of Meta’s handling of artificial intelligence-generated content, saying the company’s existing rules are inadequate as synthetic media becomes increasingly prevalent across its platforms.

In its two latest decisions, the board examined Meta’s handling of AI-generated videos involving real people. It said both cases demonstrated shortcomings in the company’s content moderation and labelling systems.

The board said Meta’s rules are “consistently and fundamentally inadequate to address the rapid rise of AI-generated content on its platforms”.

AI Video Targeting Scottish Politician

One case concerned a realistic AI-generated video depicting a Scottish politician appearing to say: “refugees are welcome here, even if they rape our women, because white people do that too.”

The politician’s voice had been replicated in the synthetic video, which the Board said was realistic despite being artificially generated. The politician described the impact as “quite traumatic”.

Meta told the Board that it had not removed the video after it was reported because none of its “trusted partner” organisations had flagged it. The company also said the content did not appear to interfere with voting or other processes.

The video did not carry an AI-generated-content label. According to the information presented to the Board, this was because the person who uploaded it had not disclosed that AI had been used to create the video.

The Oversight Board concluded that the video should have been removed under Meta’s hateful conduct rules. It also said the content should have carried a prominent AI-generated label.

AI Manipulation Used To Ridicule Volunteer

The second case involved an AI-generated video based on a television interview with a volunteer discussing menstrual health education.

The Board said that, after the original interview, accounts across social media platforms digitally manipulated footage from the interview to ridicule the woman. The altered videos subsequently went viral.

Meta automatically closed a report about the content and a subsequent appeal, according to the Board. The video also lacked an AI-generated label.

Meta eventually removed the video for violating its bullying rules after the Oversight Board became involved.

YOU MAY ALSO LIKE: US Judge Rules Pentagon’s Anthropic Designation Was Unlawful

The Board said the two cases illustrated how AI-generated content can be used to target individuals, while inconsistent labelling can contribute to the harm caused by such material.

“From politicians to private citizens, AI-generated deepfakes are increasingly being used to harass and silence women from engaging in public discourse,” Oversight Board Co-Chair Pamela San Martin said.

“These cases demonstrate a broader, troubling pattern in which women who engage publicly on issues are disproportionately subjected to harassment and misinformation.”

Board Calls For Changes To AI Moderation

The latest decisions continue a series of criticisms from the Oversight Board, which makes policy recommendations to Meta and has previously called for changes to the company’s approach to AI-generated content and labelling.

Meta’s current AI labels typically appear as a small “AI Info” banner. The company has also not agreed to implement many of the Board’s recommendations concerning AI labelling in other recent cases.

The Board has now recommended, among other measures, that Meta prioritise likely AI-generated content for review when such material is reported.

Meta did not respond to a request for comment on the Board’s latest criticism. The company has 60 days to respond to the recommendations.