How TikTok’s New Standards Are Changing the Promotion of AI Content 

In recent years, TikTok has introduced strict rules for videos created or modified using artificial intelligence. Previously, neural networks were used freely; now, the platform requires full transparency. All videos created by AI or significantly altered by it must be marked with a special label. In addition, TikTok has learned to recognize such AI-generated content on its own, even if the creator tried to hide it. Finally, users can now customize how many AI videos they want to see in their feed, which directly impacts the reach of ad campaigns and conversion rates.

When a video must be labeled as “made by AI”

According to TikTok Support’s official guidelines, labeling is required in several cases.

First, it is necessary if the video is entirely generated by a neural network. For example, a video created in Midjourney, Runway, Pika, or Sora, or an animation entirely made by artificial intelligence. This also includes deepfakes, where the faces of real people are replaced with fictional ones.

• 

Second, labeling is required if the video was filmed with a camera but then heavily modified using AI. This includes:

• face swapping;
• realistic filters that alter facial features, movements, or body type;
• a voice entirely generated by a neural network and mimicking human speech;
• a background or scene that is artificially created and looks real.

If artificial intelligence was used only for minor processing (stabilization, quality enhancement, color correction), labeling is not required. The main principle: if a viewer might be under the illusion that they are watching a real recording, when in fact it is the work of a neural network, the video must be labeled as AI content.

How TikTok identifies AI content

Recently, the platform has been implementing technologies that automatically detect synthetic videos, even if the creator has not labeled them. You can learn more about how the platform’s algorithms work in general here: https://affcommunity.org/en/how-tiktok-algorithms-work-in-2026-what-publishers-should-consider-before-launching-an-advertising-campaign/ 

One such technology is the C2PA (Content Credentials) standard. This is an international system that adds metadata to videos:

• where the video came from;
• what tools were used;
• whether changes were made using AI.

TikTok has learned to read this data and automatically applies the “Made by AI” tag.

The second technology is watermarks. These can be invisible, embedded directly into the video’s pixels: they are invisible to the naked eye, but TikTok’s algorithms easily recognize such watermarks. This type of labeling is already being tested in the US and Europe, and will soon be rolled out worldwide. Thanks to it, it becomes nearly impossible to hide the fact that a video was created by a neural network.

Users decide for themselves how much AI content they want to see

TikTok has added a new feature to its settings. Now, every user can choose how often they want to see videos marked as AI content in their feed. You can set the “less AI” mode: in this case, the algorithm will show such videos less frequently. Or the “more AI” mode, and then the feed will be filled with synthetic and experimental content.

This decision has a significant impact on advertisers and creators. Previously, reach depended solely on how interesting the video was. Now, the “content type” factor is added to this. If the target audience has selected “Less AI” mode, then ads created by neural networks (or AI-generated creativity), will be shown less frequently. The cost per impression may rise, while the number of views will automatically decrease. If, on the other hand, the audience is open to new content, such ads may, conversely, receive more impressions and higher conversion rates.

What This Means for Creators and Businesses

TikTok is effectively establishing a three-tier system for controlling AI-generated content. First, videos are checked for watermarks and C2PA metadata. Then, if the content is fully generated or heavily modified by artificial intelligence, mandatory labeling is required. Finally, users can filter out such videos using new settings, and the algorithm takes their preferences into account when distributing views.

For creators, this means that hiding the use of AI will no longer be possible. However, labeling can serve as an honest signal to viewers, showing that the creator follows the platform’s rules. For advertisers, choosing AI-generated creativity becomes part of their media strategy: they must consider how the audience perceives synthetic content, or risk losing reach and reducing conversion rates. You can read more about creativity requirements on platforms here.

Transparency and compliance with TikTok’s new standards are shifting from a recommendation to a mandatory requirement for successful promotion. All creators who want to monetize their channel on the platform should take this into account.