YouTube is making AI labels easier to read (and applying them automatically)

At this point, AI-generated content on YouTube is a fait accompli. Like it or not, AI powers many of the top channels on the world’s most prevalent video platform, and YouTube’s commitment to new AI features will only make the so-called “slop” more pervasive.

Detractors may not be able to purge AI videos from their YouTube feeds, but a recent update will make it easier to tell manmade content from machine-generated fare. YouTube’s enhancements to its AI labeling system will increase the visibility of those tags while, in many cases, applying them automatically.

YouTube introduced its AI labels in 2023 and mandated that creators append those tags to all content that was made using the assistance of AI video generators. A year later, YouTube took its enforcement a step further by adopting the C2PA standard, which provides details about videos and maintains that metadata between platforms.

Even though YouTube has taken the topic of AI disclosures seriously, its slop problem has not been solved. A purge initiated at the start of 2026 targeted some of the most notorious purveyors of low-effort, AI-generated content, but those channels pop up faster than YouTube can deal with them.

AI videos have become so pervasive that more than 200 advocacy experts

co-signed an open letter addressed to YouTube CEO Neal Mohan. The letter urged Mohan to adjust YouTube’s recommendation engine so that AI slop cannot be fed to kids.

YouTube’s latest update will help it move fast enough to stay ahead of AI slop. As described in a blog post, the “more visible, simplified labels” will appear directly below the player on long-form videos. For Shorts, the labels will appear as an overlay.

A pivot to “automatic AI detection” will make YouTube’s disclosures even more efficient. Beginning this month, YouTube will automatically apply AI labels when its systems pick up on “significant photorealistic AI use.”

Creators can override those labels via YouTube Studios, but there are a few caveats attached to that form of control. Labels cannot be removed from videos whose C2PA metadata indicates the use of AI models, and videos generated with proprietary Google AI tools like Veo will keep their labels as well.

That final restriction should feel encouraging for AI skeptics who are worried about YouTube’s active development pipeline for AI products. Yes, those products are going to keep on coming, but when they do arrive, they’ll at least (in theory) be properly labeled.