Overview
- The control appears in Manage Topics within Content Preferences and will roll out in the coming weeks with options from “see less” to “see more,” without a full block.
- TikTok says more than 1.3 billion videos on the platform carry AI-generated labels, against roughly 100 million daily uploads.
- To improve detection, TikTok is piloting an invisible watermark readable only by TikTok and will apply it to AI created with its tools and to uploads that include C2PA Content Credentials.
- The platform will keep using C2PA metadata, noting it can be lost when content is edited or reshared, and it continues to require labeling of realistic AI and bans harmful deepfakes.
- TikTok announced a $2 million AI literacy fund involving groups like Girls Who Code, as reporting highlights planned cuts to 439 London trust and safety roles and the company cites a 76% drop in graphic content seen by human moderators.