YouTube spells out what AI can't monetize
YouTube split its inauthentic-content rule into three named categories on July 16, aiming squarely at repetitive AI-made videos rather than AI use itself.
YouTube clarified its Partner Program monetization rules on July 16, breaking its existing “inauthentic content” ban into three specific categories that videos can’t fall into and still earn ad revenue. The first covers generic, repetitive, or template-driven videos, the kind AI, CGI, or a formula can churn out with little real variation between uploads. The second targets content built to distress or manipulate viewers emotionally. The third covers AI personas discussing sensitive subjects like health, finance, or legal advice.
Matt Halprin, YouTube’s trust and safety chief, framed the change as a targeting problem rather than an AI ban:
AI can actually allow people to make a lot of videos. Sometimes those videos are great… But that exact same new tool can allow you to make lots of videos really quickly that are very similar.
The policy builds on a crackdown YouTube first announced in July 2025, and the new categories give creators and reviewers a concrete checklist instead of a vague “inauthentic” standard.
For anyone running an AI-assisted channel, the read is narrow but real: volume and repetition are the trigger, not AI use on its own. A well-differentiated AI-made video should clear the bar the same way a human-made one does; a templated batch upload won’t, regardless of who or what made it. Compare this to how Deezer’s AI music detector flags synthetic tracks by pattern rather than by origin, and to the copilot-searchleak data-theft case for another instance of a platform drawing a narrower line around AI output than a blanket ban.