Meta criticized for inadequate AI deepfake moderation during Iran-Israel conflict
Meta's Oversight Board found that the company failed to properly label an AI-generated deepfake video about damage in Haifa, Israel during the 2025 Iran-Israel conflict. The board ruled that Meta's approach to detecting and labeling AI-generated content is inadequate, particularly during armed conflicts, and called for the company to overhaul its content moderation policies.
What happened
Meta's Oversight Board found on March 10, 2026, that Meta failed to label an AI-generated video claiming to show extensive damage to Haifa, Israel during the Iran-Israel conflict in June 2025. The video, which was debunked by fact-checkers, falsely depicted damage with overlaid text reading 'Live now β Haifa Towards Down' and a date.
Meta relies largely on users to self-disclose when they post AI-generated content, or waits for complaints to its moderation team before labeling such material. The Oversight Board found that this approach is 'neither robust nor comprehensive enough to contend with the scale and velocity of AI-generated content' particularly during armed conflicts.
Reported impact
The Oversight Board stated that the proliferation of AI-generated videos related to global military conflicts had 'challenged the public's ability to distinguish fabrication from fact ... risking a general distrust of all information.' A wave of viral war footage spread during the conflict, with much gaining millions of views before fact-checkers exposed it as AI-generated.
Index analysis
The Oversight Board found that Meta's handling of the Haifa video raised recurring issues about 'inefficiencies in Meta's current approach during armed conflicts.' The board said Meta should proactively label fake AI content much more frequently and implement content provenance standards at scale.
The board called for Meta to overhaul its AI rules, invest in stronger AI detection tools, and develop better labeling methods for AI-generated content. It also recommended that Meta amend its policies to ensure timely and adequate response to deceptive AI-generated output.
Governance controls
Meta said it would label the video in question within seven days following the Oversight Board's ruling. The Oversight Board frequently disagrees with Meta's content moderation rulings but the company has continued to loosen its approach to policing content.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.