In a nutshell: As deepfakes come under the spotlight following the increased abilities and availability of AI tools, Meta's Oversight Board says the company should make changes when it comes to regulating this type of content. The board says Meta's methods for identifying AI-generated content are lacking when it is created at scale, especially during times of conflict or crisis.
The latest criticism centers on a fake AI-generated video shared on Meta platforms last year that purported to show damage to buildings in the Israeli city of Haifa. It was posted by a user in the Philippines posing as a news source.
According to Meta's Oversight Board, several users reported the post, but it slipped through the cracks. Meta didn't review it, third-party fact-checkers didn't assess it, and the video remained online without a high-risk AI label until the board stepped in.
The board, whose mission is to improve how Meta treats people and communities around the world, argues that the incident exposes a larger problem with the company's current system across Facebook, Instagram, and Threads.
Right now, labeling often depends on users disclosing that AI was used or on content being escalated for special review. The Oversight Board says that approach simply isn't robust enough for the speed and scale at which AI-generated content now spreads, particularly during wars, disasters, elections, and other high-stakes events.
Among its recommendations, the board wants Meta to create a dedicated Community Standard for AI-generated content instead of relying on a patchwork of misinformation rules.
It also says Meta should apply high-risk AI labels more often, improve its automated detection systems for images, video, and audio, and clearly explain the penalties for people who fail to disclose digitally altered content.
The Oversight Board also says Meta should do more with Content Credentials, the industry framework designed to attach metadata showing where a piece of content came from and whether AI tools were involved in its creation. The group raised concerns that Meta has been inconsistent in applying those standards, including on content produced by its own AI tools.
The board previously blasted Meta's manipulated media rules as confusing and too narrowly focused on whether AI was used, rather than whether content is deceptive. Meta responded to that earlier pressure by overhauling its labeling system and replacing the "Made with AI" tag with the broader "AI info" label, though that change also drew complaints that the notices were too vague or easy to miss.
On the eve of Donald Trump's inauguration in January, Meta CEO Mark Zuckerberg announced that the company's third-party fact checkers had become too politically biased and destroyed more trust than they created. As such, they were being replaced by Community Notes.
Oversight Board urges Meta to toughen rules on AI-generated content and deepfakes
