YouTube to begin automatically detecting, labeling photorealistic AI content
![The YouTube app is displayed on an iPad in Baltimore on March 20, 2018. [AP/YONHAP]](https://img3.daumcdn.net/thumb/R658x0.q70/?fname=https://t1.daumcdn.net/news/202605/28/koreajoongangdaily/20260528114939766tpsn.jpg)
YouTube will begin automatically tagging videos that contain substantial AI content, adding a detection layer on top of the self-disclosure system that it has used since 2024, the company announced on Wednesday.
While YouTube will still require creators to manually flag their use of AI, the company's in-house system will start to identify videos with a significant amount of photorealistic AI material and apply a label to them, even when their uploader does not. The autodetection system is set to roll out this month.
“We've heard consistently from our community that they value transparency when it comes to generative AI content,” the YouTube team said in a blog post announcing the change.
The platform is also consolidating its various AI disclosures into a single label and giving it more visible placement.
For long-form videos that are detected to have been photorealistically altered or generated by AI, a label indicating AI use will now appear directly below the video player and above the description.
![A screenshot from YouTube's official blog, announcing the new AI labeling system [SCREEN CAPTURE]](https://img2.daumcdn.net/thumb/R658x0.q70/?fname=https://t1.daumcdn.net/news/202605/28/koreajoongangdaily/20260528114941251djdr.jpg)
On Shorts, YouTube's short-form vertical video service, the label will appear as an overlay on the video itself. Disclosures for animated, unrealistic or lightly altered content will continue to be located within the expanded description.
Uploaders can update their settings in YouTube Studio if they believe their content was wrongly identified as AI-generated. The label will, however, remain permanently in place for videos made with YouTube's in-house AI tools, Veo and Dream Screen, and for content whose C2PA metadata identifies it as fully AI-generated. C2PA is an industry-standard system for digitally certifying the origin and editing history of content.
The AI label does not affect how a video is recommended or whether it is eligible to earn money, according to YouTube.
This article was originally written in Korean and translated by a bilingual reporter with the help of generative AI tools. It was then edited by a native English-speaking editor. All AI-assisted translations are reviewed and refined by our newsroom. BY KIM JI-HYE [cho.yongjun1@joongang.co.kr]
Copyright © 코리아중앙데일리. 무단전재 및 재배포 금지.
- Petition calling for suspending MBC series 'Perfect Crown' gains 50,000 signatures
- Seoul concludes that Iranian-made missiles hit HMM Namu, summons Tehran's top envoy
- Starbucks Korea weekly sales sink 8 billion won in wake of 'Tank Day' controversy
- North Korea's Kim sends condolences to China's Xi as speculation over Pyongyang visit grows
- Seoul court acquits Yoon of perjury in former prime minister's insurrection trial
- Years after 'death' of mastermind of Korea's biggest Ponzi scheme, victims await justice
- Jun Ji-hyun talks silver screen return, on-set rapport with Koo Kyo-hwan and director of 'Colony'
- Korean libraries, reading spaces grow multilingual amid demographic shift, tourism boom
- As Korea's currency weakens to 17-year low, concerns mount over inflation, business pressure
- YouTube to begin automatically detecting, labeling photorealistic AI content