Skip to content
AI Primer
update

YouTube introduces automatic AI labels for undisclosed photorealistic videos

YouTube moved AI labels to more visible placements and began auto-applying them to undisclosed photorealistic AI videos. Watch Studio appeals and permanent Veo labels as Sora’s shutdown narrows consumer AI video outlets.

4 min read
YouTube introduces automatic AI labels for undisclosed photorealistic videos
YouTube introduces automatic AI labels for undisclosed photorealistic videos

TL;DR

You can read YouTube's announcement, check the exact examples in its disclosure policy, and see how the viewer-facing provenance layer works in the "How this content was made" docs. On the other side of the market, OpenAI's Help Center note puts dates on Sora's exit, while the Sora HN discussion summary captures the immediate question creators had, which was how long they'd have to preserve work.

Automatic labels

YouTube's main change is enforcement. YouTube's blog post says creator self-disclosure is no longer the only path, because the platform will auto-apply labels when its systems detect undisclosed, significant photorealistic AI content.

YouTube Updates AI Labeling for Enhanced Transparency and Creator Control

YouTube has introduced updates to its AI-generated content labeling system to enhance transparency and improve the user experience. The platform is moving labels for photorealistic and meaningfully altered AI content to a more prominent position for better viewer visibility. Additionally, YouTube is rolling out automated detection systems that will apply labels to videos if a creator does not disclose AI usage but the system identifies significant photorealistic AI content. Creators retain the ability to challenge automated labels in YouTube Studio, although content created with YouTube's own AI tools, such as Veo or Dream Screen, will have permanent disclosures. These updates do not affect video recommendations or monetization eligibility.

That label does not change recommendations or monetization on its own, according to YouTube's announcement. Creators can still contest the call in Studio, which is the practical control the HN core summary highlighted.

Label placement and scope

The placement shift is simple but material for anyone posting AI video. YouTube's announcement moves long-form disclosures directly below the player, puts them as overlays on Shorts, and leaves unrealistic, animated, or slightly altered content in the expanded description.

YouTube to automatically label AI-generated videos

For creators, this is a platform-policy change that could make AI use more visible to viewers and may affect how you disclose tool-assisted or fully generated videos. The practical questions raised in the thread are less about model internals and more about false positives, appeal friction, and whether AI-heavy channels can still dodge labels by editing or staging uploads.

YouTube's disclosure policy gives the three clearest examples of content that must be labeled:

  • a real person made to say or do something they did not do
  • altered footage of a real place or event
  • a realistic scene that never happened

The same help page also carves out a lot of ordinary creator workflow. Beauty filters, color correction, clearly fantastical scenes, and other non-misleading edits do not require disclosure.

False positives, evasion, and the gray zone

The community response went straight to detector quality. According to the HN discussion roundup, commenters worried about false positives, YouTube's existing moderation miss rate, and whether bad actors will just edit around the trigger.

Discussion around YouTube to automatically label AI-generated videos

Thread discussion highlights: - numpad0 on false positives in AI detection: The commenter asks about the overall state of AI-based AI flagging tools and says they have seen absurd false positive rates, questioning how reliable YouTube’s automated labeling will be. - CM30 on moderation accuracy vs evasion: They note YouTube already mislabels videos in other systems and argue this could help, but only if innocent creators aren’t wrongly hit and scammers can’t easily work around it. - jongjong on AI slop in kids’ recommendations: A parent says YouTube recommended creepy AI-generated videos to their child and that the platform became unusable for that purpose.

The grayest question came from a thread comment about documentary-style AI videos, which asked whether channels built from AI scripts, synthetic voices, and generated visuals, but not obvious photorealistic fake footage, will slip through. YouTube Help's viewer disclosure docs suggest the answer may depend on how strongly the content reads as meaningfully altered reality, because the disclosure system is tied to both creator input and provenance signals like C2PA.

Sora's exit narrows consumer AI video distribution

Sora shutting down does not change YouTube's policy, but it does change the distribution map around AI video. OpenAI's discontinuation page says the Sora web and app experiences ended on April 26, 2026, while the Sora API remains until September 24.

Goodbye to Sora

For creators, the key issue is that a major AI video app is being shut down, with uncertainty around preserving work and access timelines. The thread suggests the consumer social-feed angle never became a strong enough product, even though commenters still see the underlying Sora video model as impressive.

That split matters because the consumer surface disappeared first. the HN page summary framed the shutdown as the end of a creator-facing app, while OpenAI's Help Center notice says data can still be exported, with permanent deletion coming after any final export window closes.

Discussion around Goodbye to Sora

Thread discussion highlights: - ChrisArchitect on official shutdown notice: We’re saying goodbye to the Sora app... We’ll share more soon, including timelines for the app and API and details on preserving your work. - bontaq on developer-facing video products: OpenAI is also discontinuing a version of Sora for developers and won’t support video functionality inside ChatGPT, either. - password54321 on refocus on coding and business users: OpenAI’s top executives are finalizing plans for a major strategy shift to refocus the company around coding and business users

Further reading

Discussion across the web

Where this story is being discussed, in original context.

Share on X