Meta Oversight Board Says Meta's Safeguards Inadequate To Address AI-Generated Content On Its Platforms
META•Source note
Further company coverage: META.O
Oversight Board finds safeguards inadequate
Sept. 17 (Reuters) - Meta Oversight Board:
- Meta Oversight Board: Meta's safeguards consistently inadequate to address rise of AI-generated content on its platforms
- Meta Oversight Board: ordered Meta to remove AI-generated deepfakes targeting women
- Meta Oversight Board: recommends 9 policy changes to Meta including automated deepfake removal, stronger AI content labeling
- Meta Oversight Board: recommends Meta ensure policies prohibiting voting, census interference are globally applicable and not limited to U.S.
- Meta Oversight Board: recommends Meta should prioritize human review of deepfake harassment reports using AI signals
- Meta Oversight Board: recommends Meta should publish annual data in Transparency Center on number of times it applies AI labels to content
- Meta Oversight Board: urges Meta to demonetize, demote or remove "high risk" AI content from user recommendations with escalating penalties for repeat-offender accounts
- Meta Oversight Board: recommends Meta should expand "unwanted manipulated imagery" policy to cover deepfakes of private individuals
- Meta Oversight Board: orders Meta to remove deepfake videos of a UK local politician for hate speech and an AI-generated video of a female campaigner
