TikTok announced the addition of an "AI-generated content (AIGC)" slider within the "Manage Topics" tool, allowing users to adjust the proportion of AI videos in their For You feed - four options from "less" to "more" are available. The system will then reorder the recommendation pool accordingly, rather than blocking it entirely. This feature will be rolled out globally in the coming weeks and is seen as a response to Meta's "Vibes" pure AI video stream and OpenAI's social platform Sora.

TikTok also began testing "invisible watermark" technology, embedding hidden markers that only TikTok can read into content created by its AI tools (AI Editor Pro) and uploaded content with C2PA Content Credentials, to prevent loss of metadata after re-uploading or editing. The company said that the invisible watermark will run in parallel with existing visible labels and C2PA metadata, forming a three-layer identification system.
In addition, TikTok established a $2 million "AI Literacy Fund," collaborating with organizations such as Girls Who Code to create educational content, teaching teenagers to identify deepfakes and understand algorithmic recommendation mechanisms. The first courses are expected to launch in spring 2026.
