TikTok states that some anti-Semitic content approvals occurred after human reviewers overrode guidance produced by its AI systems. Reporting on the platform’s conduct is drawn from testimony or evidence presented at a hate-related inquiry, where TikTok explains how moderation decisions can vary between automated systems and human review. TikTok says that, in at least some instances, its AI may have recommended rejecting or removing the content, but the final decision was made by human reviewers who overruled that recommendation.
Both outlets say TikTok removes thousands of posts as part of its enforcement actions, while the inquiry focuses on how TikTok’s moderation tools operate and how errors can arise. The reports also indicate that TikTok is examining the circumstances in which AI and human judgment diverge, and that the hearing includes discussion of whether AI can sometimes make better decisions than people. The accounts describe a moderation process involving both AI systems and human review rather than relying exclusively on either approach.
