TikTok Labels 1.3 Billion AI Videos, Tests Feed Controls

The platform is rolling out sliders that let users dial down synthetic content in their For You page, but finding them requires digging through settings.

Scroll through TikTok right now and you'll see AI content dominate the feed. The platform has already labeled 1.3 billion videos as synthetic according to AP News, and that number only captures content creators bothered to identify. Now TikTok is testing giving users control over those feeds

The new Manage Topics sliders, buried in Content Preferences settings, let users adjust their exposure to AI-generated videos rather than blocking them entirely. According to INQUIRER.net, the controls aren't widely available yet. Most users checking their settings today won't find them. This measured rollout suggests TikTok is treating AI content management as a calibration problem rather than an on-off switch, acknowledging that synthetic media has already become integral to the platform's ecosystem.

Those 1.3 billion labeled videos however represent only a portion of total AI content on TikTok. The platform's 2026 policy updates mandate clear labeling for all realistic AI-generated content, according to Darkroom Agency, but enforcement depends on self-reporting or detection systems catching unmarked synthetic media. The gap between what's labeled and what exists creates what researchers call an attribution void.

TikTok's approach differs from Pinterest's recent tuners that limit AI content in specific categories. The TikTok sliders appear to work across all content types, adjusting the algorithm's preference weighting rather than filtering by category. PodcastVideos.com reports the platform is also developing invisible watermarking that persists through editing, suggesting TikTok recognizes that current labeling systems fail when creators crop, filter, or repost content.

The timing follows mounting creator frustration. Video professionals have watched AI-generated content achieve viral reach using techniques that would get human creators banned: manipulated thumbnails, synthetic voices mimicking celebrities, entire channels built on automated content farms. The new controls acknowledge this tension without fully shutting synethic content altogether. For now, some users can reduce their exposure to labeled AI content, but unlabeled synthetic media will still flow through.

TikTok is coupling these controls with a $2 million AI literacy fund, suggesting the company sees user education as essential to the rollout. But literacy initiatives face the same detection problem as the controls themselves. Teaching users to spot AI content becomes less effective as generation quality improves and creators get better at hiding synthetic origins.

TikTok hasn't announced when the controls will reach all users, and the company declined to provide a timeline to the outlets covering the rollout. The platform appears to be watching engagement metrics from test users before committing to full deployment.

The real test comes when these controls go wide. Will users actually dial down AI content when given the option? Or will the sliders reveal that audiences have already accepted synthetic media as entertainment, regardless of origin? The answer might determine whether platforms continue investing in transparency tools or abandon them altogether in an effort to maintain platform engagement.