A new way to tame the AI flood on TikTok
TikTok is testing a new control that lets you tell the app you want to see less AI generated content in your feed. It is not a full AI off switch, but it is the clearest sign yet that the platform knows people are getting tired of low effort AI spam clogging the For You page.
The update was announced at TikTok’s annual European trust and safety forum in Dublin. In the next few weeks, the app’s existing manage topics section will get a new toggle focused on AI. Turn it on and TikTok says it will help dial things down if you are not into AI heavy videos.
There is an important catch though. TikTok describes this as a way to tailor your feed, not to completely remove AI from it. In other words, you can ask to see less AI content, but you cannot fully opt out. AI will still be in the mix, just not as aggressively if the system works as promised.
This new control sits on top of TikTok’s growing AI detection system. The company says it has already identified around 1 point 3 billion AI generated videos posted to the platform. That sounds huge, but it is still only a fraction of what is uploaded overall. According to The Guardian, more than 100 million videos hit TikTok every single day, so there is likely a lot more AI content flying under the radar.
Invisible watermarks and AI labels
To support the new toggle, TikTok is also adding a more technical feature in the background. The platform is rolling out what it calls invisible watermarking to help label AI generated content even when it gets downloaded, reedited, and reuploaded by other users or on other platforms.
Right now TikTok already requires people to label realistic AI generated videos. A lot of this happens through C2PA Content Credentials. Think of these as a kind of ID tag baked into the file that says how the content was created and whether AI was involved. TikTok uses this metadata to spot AI images and videos at scale, which is part of how it reached that 1 point 3 billion detection count.
The problem is that these labels are easy to strip. If someone downloads a video, crops it, or runs it through another editor, those credentials can disappear. TikTok openly admits this is a common industry challenge for every social platform trying to keep track of AI content.
The new invisible watermarking tool is meant to fix that. Instead of relying only on a label you can see and delete, TikTok will quietly embed a signal into videos created with its own AI powered editing tools and with content that already carries C2PA credentials. This signal should survive most basic edits and reuploads, making it much harder for AI clips to pretend they are fully human made.
For normal users, you will probably never see this watermark. It is not a logo on the corner of your screen. It is more like a hidden fingerprint that TikTok’s systems can scan to work out whether a video is AI generated and then adjust recommendations or apply labels accordingly.
Why TikTok suddenly cares about AI literacy
Publicly, TikTok is trying to stay upbeat about generative AI. In its press material the company says it wants to safeguard and empower positive experiences with AI. It avoids calling AI spam a problem directly. But if you look at what it is actually doing, the message is clear. AI content has become big enough and messy enough that it is now a trust and safety issue, not just a fun toy.
Alongside the new AI controls TikTok is launching a 2 million dollar AI literacy fund. The money will go to organisations such as Girls Who Code, with the goal of creating TikTok content that teaches people about AI, how it works, and how to stay safe around it.
That kind of investment is not something you make if everyone already understands AI and uses it responsibly. It is more of a damage control move. There are concerns about fake videos, misinformation, and people not realising when what they are watching is generated rather than recorded.
At the same time TikTok has drawn criticism for something that looks like the opposite of safety. The company has plans to cut hundreds of jobs in its trust and safety teams, reportedly around 439 roles. Trade unions and online safety experts have attacked the move, saying it weakens human moderation just when the platform most needs it.
TikTok’s response is that this is a reorganisation and that it wants to evolve moderation using new technology. In other words, lean more on AI monitoring systems and less on human moderators. That is a tricky sell when AI content is already causing headaches for users and regulators.
Put together, the story looks like this. AI tools make it cheap and easy to pump out endless videos. TikTok benefits from that flood of content, but also has to manage the chaos it creates. So it is building better detectors, invisible watermarks, and user controls that let you resist the worst of the AI slop without banning it entirely.
If you use TikTok regularly, the important takeaway is simple. Soon you should get a new toggle in the manage topics menu that lets you tone down how much AI generated stuff shows up in your feed. It will not magically bring back a fully human For You page, but it is a welcome bit of power in a feed that is usually shaped mostly by the algorithm.
Original article and image: https://www.pcgamer.com/hardware/inundated-with-slop-tiktok-tests-feature-that-will-let-users-request-to-see-less-ai-generated-content-in-their-feeds/
