Particle.news
Download on the App Store

TikTok Tests Slider to Dial AI Videos Up or Down, Adds Invisible Watermarks

The move aims to let viewers curb synthetic clips using a new control backed by stronger provenance signals.

Overview

  • The AI-generated content setting lives in Manage Topics and will roll out in the coming weeks, letting users request more or less AI in the For You feed without removing it entirely.
  • The control is opt-in and does not affect other parts of TikTok like the Following feed, so AI videos may still appear even when users choose to see less.
  • TikTok will apply an invisible watermark to AI videos made with its tools and to uploads carrying C2PA Content Credentials to make labeling harder to strip during reuploads or edits.
  • The company says about 1.3 billion videos on the platform currently carry AI-generated labels, reflecting both automated detection and creator disclosures with known limits.
  • TikTok announced a $2 million AI literacy fund involving groups such as Girls Who Code and highlighted the changes at its Dublin trust and safety forum as it pursues more automated moderation alongside planned cuts to 439 London trust-and-safety roles.