TikTok adds AI content slider: why giving users a dial, not just labels, matters
TikTok has added a slider control for AI-generated content to its Manage Topics panel, letting users adjust how much synthetic material appears in their For You feed. The slider rolled out in TikTok's existing "Manage Topics" tool, which already allows users to tune categories like dance, sports, and food.
The feature matters because it is the first time a major platform has given audiences direct, graduated control over AI content, rather than just applying labels. TikTok clarified that the setting "is intended to help people tailor the diverse range of content in their feed, rather than removing or replacing content in feeds entirely".
Why TikTok built a dial, not a toggle
Users cannot turn AI content off entirely. The slider reduces exposure but does not enforce a hard filter. TikTok will not allow the slider to remove AI content completely, and it influences recommendations based on user behaviour rather than enforcing a hard filter.
The slider only affects the For You Page, so users may still see AI clips through search or the Following tab. Changes appear gradually as the algorithm retrains.
Estimates suggest over 1.3 billion AI-generated clips already exist on the platform, a volume that makes binary on/off controls impractical. The slider reflects that reality.
How it works
Users navigate to Settings and privacy, then Content and display, then Content preferences, then Manage topics. The AI-generated content slider can be moved fully left to reduce it.
The feature is gradually rolling out worldwide over the next few weeks, so not all users will see it immediately.
The tool only works on properly labelled videos, so heavily edited clips or externally generated videos can confuse detection. Invisible watermarks are proprietary, so only TikTok can read them, which limits transparency.
TikTok couples control with stronger detection
The slider launched alongside upgrades to TikTok's labelling infrastructure. TikTok will start adding invisible watermarks to AI-generated content made with TikTok tools like AI Editor Pro, and is also adding them to content uploaded with C2PA's Content Credentials.
These new watermarks can only be read by TikTok, making them harder to remove and easily traceable regardless of any edits being made. The move addresses a persistent problem: the label often gets lost once AI videos are uploaded or edited elsewhere.
Creators are still required to disclose AI content. TikTok requires creators to label realistic AI-generated images, audio, or video, particularly when they depict real people or could be confused with authentic recordings.
Like those controls, the AIGC setting is intended to help people tailor the diverse range of content in their feed, rather than removing or replacing content in feeds entirely.
TikTok blog post, November 2025What other platforms are doing (and not doing)
Meta labels AI-generated content across Instagram, Facebook, and Threads using C2PA signals. TikTok requires creators to disclose AI content. But these are labels, not controls.
Pinterest added tools that let users limit how much AI-generated content they see on the platform, and users can now personalize their feeds to restrict generative AI imagery in select categories. The tools are available in the app's Settings menu, under Refine your recommendations, where users can configure whether they'd like to see less GenAI content in certain categories prone to AI-generated imagery, like beauty, art, fashion, and home décor.
Pinterest rolled this out in October 2025, a month before TikTok. But TikTok's implementation is broader: it applies platform-wide, not category by category.
YouTube will now automatically label videos that use significant photorealistic AI, and starting in May 2026, YouTube will use new internal signals to help identify AI-generated content and label it accordingly. But YouTube has not launched user-facing controls to reduce AI content in feeds.
The pattern is clear. Labels are table stakes. TikTok and Pinterest have moved to the next layer: giving audiences a way to act on the information.
The context: AI content is flooding every feed
GenAI content now makes up 57% of all online material, according to academic literature cited by Pinterest. The same figure appears across multiple platform announcements.
By November, TikTok had already labelled 1.3 billion videos as AI-generated. On TikTok, 294 of 500 For You videos shown to a fresh account were identified as AI slop, according to research by Kapwing. On YouTube, 104 of the first 500 Shorts, or 21%, were AI slop.
The Kids category on TikTok showed the highest AI density. Of the 2,000 videos Kapwing reviewed in TikTok's Kids category, 57% were AI slop.
The volume makes TikTok's slider less a feature than a necessity. For many users, the question is no longer whether they will see AI content, but how much.
What this means for social marketers
The shift from labels to dials changes the calculation for creators and brands using AI tools.
Labelling is now the floor. Every platform expects it. Missing it risks strikes, not just algorithmic down-ranking. Enforcement of unlabelled AI content on TikTok increased 340% compared to 2024, with 51,618 synthetic media removals in the second half of 2025.
User tolerance is the new constraint. If a significant portion of your audience has dialled AI content down, unlabelled or poorly disclosed AI work will reach fewer people, even if it is technically compliant. The slider does not block content outright, but it does reduce distribution.
Authenticity is a competitive position. If half the feed is synthetic, being verifiably human becomes a point of difference. Lean into it.
Platform tools leave a trail. TikTok will start adding invisible watermarks to AI-generated content made with TikTok tools like AI Editor Pro. If you create with platform-native AI features, assume the platform knows, labels it, and surfaces that to users who ask.
The commercial upside: user-controlled feeds are more engaged feeds. If people can reduce what they do not want, they are more likely to stay and scroll. That makes the remaining audience more valuable, not less.
The bigger picture: user agency over algorithmic opacity
TikTok's move reflects a broader shift across social platforms. Social platforms are handing over some of their power by allowing users to personalise their algorithms with the help of AI, and apps like Threads, Instagram, and TikTok have begun introducing tools that let people train their own algorithms and influence what appears in their feeds.
The Manage Topics panel, launched in June 2024, was the foundation. Manage Topics, available via Settings and Content Preferences, enables users to indicate how interested, or not, they are in different, broad-ranging topic categories, and TikTok launched this option with some users in August last year and has been testing it ever since.
Adding AI content as a topic within that framework signals how platforms now view synthetic media: not as a category to ban, but as a user preference to manage.
The question is whether other platforms follow. YouTube has the detection infrastructure. Meta has the labelling. Neither has shipped the dial.
For now, TikTok owns the user-controlled AI feed. That matters commercially. It is also the shape of what is coming everywhere else.