YouTube is moving from a creator-driven disclosure model to a hybrid one. The platform has begun automatically labeling videos that appear to contain significant photorealistic AI generation, even when the uploader has not self-disclosed the use of generative tools.
Starting in May, YouTube says it is using new internal signals to flag AI-generated material and apply the label itself, while still expecting creators to disclose their use of AI. The change makes the disclosure harder to miss for anyone encountering photorealistic, AI-altered, or AI-generated material, and it does not affect recommendations or monetization.
How does the new detection system work?
YouTube’s policy requiring creators to flag AI content has been in place for more than two years. It was introduced alongside a disclosure tool inside Creator Studio, which asked creators to mark videos that could be mistaken for depicting a real person, place, or event. Obviously animated or fantastical content, such as a unicorn in a surreal landscape, was always exempt.
What has changed is enforcement. Starting in May, YouTube says it is using new internal signals to identify AI-generated material and apply the appropriate label itself. Creators are still expected to disclose their use of AI, but the platform will now step in when they do not.
The automatic detection roll-out follows Google’s release of Gemini Omni at Google I/O, a family of multimodal models capable of producing high-quality video with an awareness of physics, culture, history, and science.
Which videos get labeled automatically?
Labels will be applied automatically in two main scenarios:
- When YouTube’s internal systems flag content as containing significant photorealistic AI.
- When uploaded files carry C2PA metadata indicating the content was fully AI-generated. This metadata is permanent and travels with the file.
YouTube notes that labels cannot be removed from videos produced with its own generative tools, including Veo and Dream Screen. For other videos that are misidentified, creators can update the disclosure status, though the platform retains the ability to override those edits when the content originated from YouTube’s tools.
Where will the new labels appear?
The labels themselves are also being redesigned. Previously, AI labels appeared in the expanded description, with a more prominent on-player label reserved for sensitive categories such as health or news. Now, the label sits directly below the video player and above the description on long-form uploads, and appears as an overlay on YouTube Shorts.
For content that is only slightly altered, animated, or unrealistic, the label will continue to live in the expanded description. YouTube’s stated goal is to make the disclosure harder to miss for anyone encountering photorealistic, AI-altered, or AI-generated material.
Will AI labels affect recommendations or monetization?
YouTube has clarified that receiving an AI label will not affect how a video is recommended, nor will it alter monetization. The change is positioned as a transparency measure rather than a punitive one.
What other AI work is YouTube doing?
The automatic labeling feature arrives alongside other AI work at the company. YouTube has recently expanded its deepfake face-matching detection to any adult user, after earlier tests with public figures, politicians, and creators. The platform has also been rolling out AI-driven features including an interactive search experience called Ask YouTube, playlist generation for YouTube Music, and AI video summaries.
What is the C2PA standard?
C2PA is a content provenance standard that embeds cryptographically signed metadata describing how a piece of media was created and edited. OpenAI recently committed to the standard, joining Nvidia, Kakao, and ElevenLabs. YouTube’s reliance on this metadata means that provenance information attached at creation can travel with the file and surface as a label on the platform without requiring YouTube to perform its own detection.
FAQ
When did YouTube start automatically labeling AI-generated videos?
YouTube says it began automatically labeling videos containing significant photorealistic AI generation in May, using new internal signals and C2PA metadata.
Can creators remove an AI label from their video?
No, not if the video was produced with YouTube’s own generative tools, including Veo and Dream Screen. For other misidentified videos, creators can update the disclosure status, but YouTube can override those edits when the content originated from its tools.
Will an automatic AI label hurt a video’s reach or revenue?
No. YouTube has clarified that receiving an AI label does not affect how a video is recommended or how it is monetized, and the change is described as a transparency measure rather than a punitive one.
Related coverage
Want help with this in Myrtle Beach or Long Island? Talk to us.
NewSunSEO
Grow in Google and AI Search
AI-SEO, local search, paid ads and websites from one team, with offices on Long Island and in Myrtle Beach.
- AI-SEO
- Local AI-SEO
- AI Consulting
- PPC
- Web Design
- Web Development
- Long Island Office(631) 320-1700
- Myrtle Beach Office(843) 492-7131
Request a Free Proposal
Tell us about your business and where you want to grow.
We will reply by email to talk through your goals and what a proposal would include.
We use your details only to reply to you and send what you asked for. See our Privacy Policy.
