Overview
- A new AIGC slider inside Manage Topics lets users choose to see fewer or more AI‑generated videos in the For You feed, with a gradual rollout in the coming weeks.
- TikTok will begin testing an invisible watermark readable only by the platform to tag realistic AI content, making removal by third parties more difficult.
- The watermark will apply to videos created with tools such as AI Editor Pro and to clips that already carry C2PA Content Credentials when uploaded.
- TikTok says it has labeled more than 1.3 billion AI‑generated videos so far, noting that labels and C2PA metadata can be lost when content is edited or reposted outside the app.
- The company reports 86% of violating posts were removed automatically in the first half of 2025 and more than 70 million fake TikTok Shop listings were deleted as AI‑enabled scams rose about 40% versus late 2024; a $2 million AI literacy fund is underway.