Meta Oversight Board Demands Dedicated AI Content Rules, Better Detection Tools
Key Facts
- The Oversight Board urged Meta to create a standalone policy for AI-generated content separate from its existing misinformation rules.
- Recommendations include investing in more reliable AI detection technology for video, audio and images, and consistently applying digital watermarks.
- The case centers on a deceptive AI-generated video claiming to show damaged buildings in Haifa during the 2025 Israel-Iran conflict that amassed over 700,000 views before Meta declined to label it as high-risk AI content.
- The board criticized Meta’s current “AI Info” labels as insufficient for the scale of AI-generated content, especially during conflicts and crises.
- Meta has 60 days to formally respond to the recommendations and previously updated its labeling approach in May 2024 following earlier board criticism.
Lead paragraph
The Meta Oversight Board is calling for sweeping changes to how the company handles AI-generated content, saying the social media giant needs a dedicated policy separate from its misinformation rules and must invest in better detection tools. The recommendations follow the board’s review of an AI-generated video falsely depicting damage in the Israeli city of Haifa during the Israel-Iran conflict that gained more than 700,000 views on Meta’s platforms. According to the board’s decision, Meta’s current approach relying heavily on user self-disclosure and infrequent escalated reviews is “neither robust nor comprehensive enough” to address the rapid spread of deceptive AI content, particularly during armed conflicts.
Background of the Case
The disputed content was posted last year by an account claiming to be a news outlet but operated by an individual in the Philippines. The video purported to show buildings damaged in Haifa amid the 2025 Israel-Iran conflict. When users reported the clip, Meta chose not to remove it or apply a “high risk” AI label that would have alerted viewers the content was created or altered with artificial intelligence.
The Oversight Board overturned Meta’s decision not to add the high-risk label. It noted that the company eventually disabled three accounts linked to the page after the board highlighted “obvious signals of deception.” The case, the board said, exposed multiple weaknesses in Meta’s current policies for handling synthetic media.
Board’s Key Recommendations
In its decision, the Oversight Board made several concrete recommendations. The primary suggestion is for Meta to establish a distinct rule specifically for AI-generated content. This new policy should clearly outline when and how users must label AI-created material and detail the penalties for violations.
The board was sharply critical of Meta’s existing “AI Info” labeling system. It argued that the approach, which depends largely on creators voluntarily disclosing AI use and on rare manual reviews, cannot keep pace with the volume and speed of AI-generated content, especially during periods of conflict or public crisis.
The group also called on Meta to develop and deploy more sophisticated detection technologies capable of reliably identifying AI-generated media across video, audio and still images. Additionally, the board expressed concern over reports that Meta is “inconsistently implementing” digital watermarks on content created by its own AI tools, such as those available through Meta AI.
Broader Criticism of Meta’s Approach
This is not the first time the Oversight Board has criticized Meta’s handling of manipulated and AI-generated media. The board has previously described the company’s manipulated media rules as “incoherent” on multiple occasions. It has also faulted Meta for over-relying on third-party fact-checking organizations to identify problematic content.
In this case, the board highlighted feedback from trusted partners indicating that Meta has become “less responsive to outreach and concerns,” partly due to reductions in internal team capacity. The Oversight Board stated that Meta “should be capable of conducting such assessments of harm itself, rather than rely solely on partners reaching out to them during an armed conflict.”
The decision also touched on the wider industry challenge. While focused on Meta, the board noted that “the industry needs coherence in helping users distinguish deceptive AI-generated content and platforms should address abusive accounts and pages sharing such output.”
Context of Recent Policy Changes
Meta has already made some adjustments to its AI labeling practices. In April 2024, the company acknowledged that its previous approach was too narrow, covering only videos altered to make someone appear to say something they did not. Following an earlier Oversight Board recommendation from February 2024, Meta expanded labeling to a wider range of video, audio and image content.
Beginning in May 2024, Meta started applying “Made with AI” labels to more types of synthetic media and began providing additional context for high-risk material, particularly political content. These changes were intended to help users better understand when AI was used to create or alter content.
However, the Oversight Board’s latest decision suggests these updates have not gone far enough to address the proliferation of deceptive AI content, especially when spread by inauthentic networks during sensitive geopolitical events.
Impact on Users, Platforms and the Industry
The recommendations come at a time of heightened concern about AI-generated misinformation. Since the start of recent US and Israeli strikes on Iran earlier this month, there has been a documented surge in viral AI-generated content across social media platforms. This includes misleading videos, images and audio that can rapidly influence public perception during armed conflicts.
For users, clearer labeling and more reliable detection tools could make it easier to distinguish between authentic and synthetic media. For Meta, implementing the board’s suggestions would likely require significant investment in technical infrastructure and policy development. The company’s 60-day window to respond formally will be closely watched by regulators, civil society groups and competitors.
The Oversight Board’s continued scrutiny also highlights the unique role of this independent body, which has powers to review specific content decisions and issue broader policy recommendations. While its decisions are not legally binding, they carry significant influence on Meta’s public image and future policy direction.
What’s Next
Meta has not yet issued a public response to the latest Oversight Board decision. The company is expected to provide a formal reply within 60 days. Given its previous adjustments following earlier board criticism, it is likely Meta will announce further policy refinements, though the extent of those changes remains to be seen.
The board has previously expressed interest in collaborating with generative AI companies more broadly. Its latest statement suggests a desire for greater industry-wide coordination on standards for labeling and addressing deceptive AI content.
As generative AI tools become more accessible and sophisticated, the pressure on major platforms like Meta to develop coherent, scalable solutions for identifying and labeling synthetic media is only expected to grow. The Haifa video case serves as a concrete example of how quickly misleading AI content can spread during international crises and why current systems remain inadequate.
Sources