YouTube adds automatic AI labels and Studio appeals for undisclosed photorealistic videos

• YouTube moved AI disclosures for photorealistic or meaningfully AI-altered videos into harder-to-miss spots, with long-form labels now sitting below the player and Shorts labels appearing as an overlay, according to YouTube's update summary and the official YouTube blog post.
• If a creator skips self-labeling and YouTube's systems detect significant photorealistic AI use, the platform can now auto-apply the label, as the main HN summary notes and YouTube details in its policy update.
• Creators can usually challenge an incorrect label inside YouTube Studio, but content made with Veo, Dream Screen, C2PA-tagged generative media, or labels applied after manual review can stay locked, per YouTube's update summary and the Help Center rules.
• YouTube says the label alone does not change recommendations or monetization, while the HN discussion roundup shows commenters immediately jumping to false positives, evasions, and viewer-side filtering as the real unresolved problems.

You can read the official announcement, the full creator disclosure policy, and the viewer-facing "How this content was made" explainer. The interesting wrinkle is that YouTube paired a transparency update with quiet enforcement teeth: auto-labels, Studio disputes, permanent labels for some AI sources, and possible penalties when creators repeatedly skip disclosure.

The biggest visible change is placement. In the official blog post, YouTube says photorealistic AI labels now sit directly below the player on long-form videos, while Shorts get the label as an on-video overlay.

Less realistic cases still get pushed down a layer. The viewer-facing Help page says non-photorealistic or animated AI disclosures can appear in the expanded description under "How this content was made."

That split gives YouTube a two-tier system:

• Photorealistic or meaningfully altered realistic video: label near the player
• Shorts with the same class of content: overlay on the video
• Unrealistic, animated, or lightly altered content: disclosure in the expanded description

The enforcement change is simple. If creators do not tick the AI-use field, YouTube can now add the label itself when its internal signals detect significant photorealistic AI use, according to the main HN summary and the official announcement.

The creator policy page makes the Studio workflow more explicit than the blog post does. Creators disclose during upload in the Attributes section, and if YouTube gets it wrong, they can usually switch the AI disclosure back to No inside YouTube Studio.

The policy also spells out what triggers disclosure in the first place:

• A real person made to appear to say or do something they did not do
• Footage of a real event or place altered with AI
• A realistic scene generated that never actually happened

YouTube also adds one notable content example that goes beyond fake video. Its Help docs list AI-generated music as content creators need to disclose when it appears realistic, which widens this beyond deepfake-style clips into audio-led creator workflows too.

The community response moved past the label almost immediately. In the discussion roundup, one commenter called the visibility boost a net positive for less technical viewers, another worried the detection stack would be wrong often enough to hurt legitimate creators, and a third suggested a browser extension that crowdsources AI-heavy channels and hides them from feeds.

That same HN thread pushed on a gap YouTube did not solve here: labeling is not filtering. The announcement says labels do not affect recommendations or monetization, while commenters kept asking for tools that separate human-made work from synthetic uploads, especially in feeds and kid-facing viewing.

One parent anecdote in the main HN thread described repetitive AI-generated videos surfacing to a 4-year-old. That is a different problem than disclosure accuracy, and it explains why creators and viewers are already talking about classification, ranking, and channel-level filtering instead of just the badge itself.

The hardest edges are buried in the docs. The YouTube blog post says disclosures stay permanent for content made with YouTube's own AI tools, including Veo and Dream Screen, and for content carrying C2PA metadata that indicates fully generative AI.

The creator policy page adds a third locked bucket the blog only hints at: labels applied after manual review cannot be adjusted in Studio. It also says creators who repeatedly fail to disclose can face manual labeling, content removal, or suspension from the YouTube Partner Program.

That leaves three separate non-appealable paths:

• Content made with YouTube's own GenAI tools
• Content containing C2PA metadata
• Content labeled after manual review

Those details turn the update from a UI cleanup into a stricter disclosure system. The front end change is the badge placement. The back end change is that YouTube now has more ways to attach the label, fewer ways to remove it in some cases, and an explicit penalty ladder when creators keep skipping the form.