Select what AI did in your video. Every use gets its own verdict, with the YouTube page behind it.
This checker needs JavaScript to run.
Choosing what AI did
The cards are grouped by what the AI touched: voice and audio, real people, places and events, other visuals, and text and production. Pick every use that applies to the finished video, not only the main one. A card shows its verdict as soon as you select it, together with its basis: either YouTube lists that exact case, or it follows from one of the three criteria in YouTube Help. A typical faceless-channel workflow shows how this reads: an AI-assisted script, a clone of your own voice, an AI thumbnail and AI-generated music give one match out of four, and the match is the music, because YouTube lists AI-generated music among its disclosure examples while the script, the thumbnail and the voice clone sit on its not-required list.
One question for unclassified cases
A few uses appear in neither of YouTube’s lists, for example a synthetic narrator that imitates nobody, AI sound effects, generic photorealistic b-roll or a synthetic presenter. These cards show Needs context and ask one question tied to a published criterion. Answering yes moves the card to Disclose and names the criterion, answering no moves it to No trigger found with the note that no published criterion matched, and Not sure leaves the card open.
The stamp and the trail
The stamp summarises the whole selection as Disclosure needed, No trigger found or Needs context. Below it, the trail lists each selected use with its basis, the reason and a link to the YouTube page behind it, and the cross on a row removes that use so you can see which single choice changed the outcome. The tally counts matches out of selected uses. YouTube says its examples are not exhaustive, so the result maps your selections to published guidance and does not replace the choice you make in Studio.
Where viewers may see the label
The preview switches between long-form and Shorts, and between photorealistic and animated or minor content. For photorealistic or meaningfully AI-altered content, YouTube places the label below the player on long-form videos and over the video on Shorts, while unrealistic, animated or slightly altered content is disclosed in the expanded description. The preview is an explainer that does not depend on your result, and its label styling is an illustration rather than a copy of the YouTube interface.
Automatic labels
The two switches cover cases where YouTube can apply a label on its own: content made with its own AI tools, and files whose Content Credentials mark the whole video as AI-made. YouTube states that some automatic labels can’t be removed, which includes these cases and labels applied after manual review, while labels based on its own detection can be corrected in Studio in most cases.
Copying the result
Copy result puts one summary sentence on your clipboard with the date of the check. Copy breakdown adds a line per selected use with its verdict, which suits a team chat or a production log.
What this checker leaves out
The checker covers the AI use disclosure only. Whether a channel can earn money is decided by separate policies, including inauthentic content, which was called repetitious content until July 2025, and according to YouTube a disclosure label on its own does not change how a video is recommended or whether it can earn money. The rules were read from YouTube Help and the YouTube Blog on September 21, 2026, and this page covers platform rules only, not legal advice.