A powerful oversight group is warning that AI-generated videos are spreading too easily on Meta’s platforms — and they’re becoming increasingly difficult for users to recognize. While the more fun and innocent AI generated cat videos flooding your feed are easy to spot, others aren’t as easy to detect.
In a new decision, the Oversight Board urged the company to strengthen its policies and detection tools for AI-generated content, particularly realistic deepfake videos that can spread misinformation during major global events.
The Oversight Board is affiliated with Meta but was set up as a "body of experts from around the world that exercises independent judgment and makes binding decisions on what content should be allowed on Facebook and Instagram."
The recommendation follows a review of a fake AI-generated video depicting destruction in Israel, which circulated online and highlighted gaps in Meta’s current moderation systems.
According to the board, Meta’s existing approach relies too heavily on users labeling AI-generated content themselves — a system that can easily fail when misleading videos go viral before moderators catch them.
Why deepfakes are getting harder to detect
Advances in generative AI have made it dramatically easier to create realistic video, audio and images that appear authentic at first glance.
Modern AI video tools can now generate footage with convincing lighting, motion and voice synchronization. As these tools improve, the line between authentic and synthetic media is becoming increasingly blurred.
That’s part of what worries experts.
When highly realistic AI videos spread during wars, disasters or elections, they can quickly shape public perception before fact-checkers have time to respond.
What the Oversight Board wants Meta to do
The board has warned that current policies were largely designed for traditional misinformation — not the new wave of highly convincing generative AI media. For that reason, it's recommending several changes designed to slow the spread of deceptive AI media:
- Create clearer rules specifically for AI-generated content
- Improve automated detection tools for deepfakes
- Label AI-generated media more clearly
- Adopt industry standards like Content Credentials, which can show whether media was created or modified using AI
Bottom line
Meta isn’t the only platform grappling with the issue. Because AI video generation tools are available for free such as Gemini, Grok and Sora, anyone with a device can now generate realistic videos in minutes.
For users, the challenge may only grow harder: the next viral video you see online could look real — even if it isn’t. If you're unsure, try using chatbots like Gemini, which can detect AI generation or dig deeper with Deep Research for the source of the video. Staying vigilant and doing your own due diligence is one of the best ways to avoid getting tricked by a deepfake.
