Meta’s Oversight Board issues nine-point deepfake policy changes

Meta's Oversight Board issued nine policy recommendations on deepfake removal, voting interference protections, and new content moderation frameworks for generative AI, following two cases including a June 23 ruling that overruled Meta's decision to keep an AI-generated sexualized video impersonating a private individual on Instagram. The Board called for a dedicated Community Standard for AI-generated content separate from Meta's legacy misinformation policy, "High Risk" labels for potentially AI-generated content, and provenance metadata embedded in outputs from Meta's own AI tools. Meta, which has historically implemented roughly 74-75% of the Board's recommendations and must typically respond within 60 days, said the nine AI-related points remain under assessment.

Meta Platforms logo public domain via Wikimedia Commons Meta’s Oversight Board issues nine-point deepfake policy changes The independent board wants AI-generated impersonations classified as non-consensual by default and calls for a dedicated community standard for synthetic content Meta https://cryptobriefing.com/markets/meta/ ’s Oversight Board handed the company nine policy recommendations targeting deepfake removal, voting interference protections, and the creation of entirely new content moderation frameworks designed for the generative AI era. The recommendations follow two high-profile cases that exposed gaps in Meta’s existing moderation playbook. One involved a sexualized deepfake video targeting a non-public figure on Instagram. The other dealt with deceptive AI content that surfaced during the Israel-Iran conflict earlier this year. The Instagram deepfake that broke the system On June 23, the Oversight Board overruled Meta’s own decision to keep a piece of content live. The content in question was an AI-generated sexualized video on Instagram that impersonated a private individual. Meta had initially decided the video could stay up. The Board disagreed, ordering its removal and establishing a broader principle: AI-generated impersonations should be treated as non-consensual by default. The reputational and societal harms from synthetic impersonation content are severe, the Board argued, particularly for women and children who are disproportionately targeted by deepfake abuse. By reclassifying these cases as inherently non-consensual, Meta would no longer need victims to come forward before taking action. AI, tech, and the markets they move—in one daily briefing. Daily. Free. Join 34,000+ readers across crypto, finance, and policy. A new rulebook for AI-generated content On March 10, the Board examined deceptive AI content that emerged during the 2025 Israel-Iran conflict, where synthetic media was being used to fabricate events and manipulate narratives. The Board found that Meta’s existing Misinformation community standard wasn’t built to handle AI-generated content at scale and called for a dedicated Community Standard specifically for AI-generated content, separate from the legacy misinformation policy. Among the specific technical measures recommended: implementing “High Risk” labels for content flagged as potentially AI-generated, and embedding provenance metadata into outputs created by Meta’s own AI tools. The Board also pushed for extending voting interference protections globally, with consistent standards applied everywhere Meta operates. Meta’s track record on compliance Meta isn’t legally required to follow the Board’s recommendations. Historically, Meta has implemented roughly 74-75% of the Board’s recommendations. Meta is typically required to respond to Board reports within 60 days. The new AI-related recommendations are still under assessment, meaning Meta hasn’t yet committed to adopting any of the nine points. Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy https://cryptobriefing.com/editorial-policy/ .