Meta’s deepfake moderation isn’t good enough, says Oversight Board
Meta’s methods for identifying deepfakes are “not robust or comprehensive enough” to handle how quickly misinformation spreads during armed conflicts like the Iran war. That’s according to the Meta Oversight Board — a semi-independent body that guides the company’s content moderation practices — which is now calling on Meta to overhaul how it surfaces and labels AI-generated content across Facebook, Instagram, and Threads.
The call for action stems from an investigation into a fake AI video of alleged damage to buildings in Israel that was shared on Meta’s social platforms last year, but the Board says its recommendations are particularly relevant right now, given the “massive military escalations” throughout the Middle East this week. In its announcement, the Board says that access to accurate, reliable information is vital to people’s safety amid the heightened risk of AI tools being used to spread misinformation.
“The Board’s findings highlight that Meta’s current system to properly label AI content is overly dependent on self-disclosure of AI usage and escalated review and does not meet the realities of today’s online environment,” the Meta Oversight Board said. “The case also highlights the challenges with cross-platform proliferation of such content, with the content appearing to have originated on TikTok before appearing on Facebook, Instagram, and X.”
Recommended steps issued by the Board include pushing Meta to improve its existing rules on misinformation to address deceptive deepfakes, and establish a new, separate community standard for AI-generated content. Meta is also being asked to develop better AI detection tools, be transparent about penalties for AI policy violations, and scale AI content labeling efforts. The latter includes ensuring that “High-Risk AI” labels are added to synthetic images and videos more frequently, and improving C2PA (otherwise known as Content Credentials) adoption so that information on AI-generated content is “clearly visible and accessible to users.”
