TikTok Tests Invisible Watermarking to Reinforce AI Content Safety
TikTok has introduced new app updates to strengthen its ability to identify and manage AI-generated content on its platform. As part of this initiative, the company is rolling out an experimental setting under its “Manage Topics” feature. Using this setting, users will be able to control how much AI-generated content they want to see in their ‘For You’ feed. The tool will also help personalize content recommendations.
To identify AI-generated videos, TikTok currently uses creator-applied labels, AI detection models, and C2PA content credentials. The app is now testing a new method called “invisible watermarking.” This hidden watermark adds an extra layer of security and can only be detected by TikTok. The watermark is difficult to remove even if the video is edited or reposted.
Over the next few weeks, this invisible watermark will be added to AI-generated videos made using TikTok’s tools—such as AI Editor Pro—as well as to content uploaded with C2PA content credentials.
The short-video platform has also announced a USD 2 million “AI Literacy Fund” aimed at helping people learn about artificial intelligence. Through this fund, institutions and experts will produce educational content explaining how AI works and how to identify AI-generated material. Collaborators from more than 12 countries will develop various tools to promote media literacy, responsible technology use, and awareness about TikTok’s features.
DBTech/KE/EI/OR
