YouTube did not ban AI video. On 16 July 2026 it did something more precise and, for anyone running a small channel, more consequential: it split its "inauthentic content" monetisation rule into three named categories, each describing a production habit rather than a tool. Generic or template-based output, emotionally manipulative clips, and synthetic personas presenting as experts on money, law or health. A channel carrying too much of any one of them cannot earn through the YouTube Partner Program. If you use AI in your workflow and your videos still carry something only you know, nothing here touches you. If your format is "same script, new topic, ship five a day," the format is the problem and it was already ineligible.
The rule is older than the headlines
The policy itself is not new. YouTube renamed "repetitious content" to "inauthentic content" on 15 July 2025 to make clear the rule covers material that is "repetitive or mass-produced," and its channel monetization policies state that such content "has always been ineligible for monetization under our existing policies, where creators are rewarded for original and authentic content." The July 2026 update is a clarification of an existing standard, which matters for two reasons: there is no grace period to wait out, and the enforcement you are seeing now is the same rule with better definitions attached.
The tool-neutral wording is worth reading closely, because it is the part most coverage flattens. The violation named on the policy page is "AI-generated content made with generic or unoriginal templates giving the impression of mass production without adding the creator's original, authentic insights or perspective." Every qualifier in that sentence is load-bearing. Generic templates. Impression of mass production. Without adding original insight. Remove any one of them and the video is not what the policy describes.
Category one: generic, repetitive or template-based
The trigger is a video that "looks like it's made with a template," and YouTube's stated standard is that "the substance of each video should be materially varied and deliver creative, educational, or other value." TechCrunch's reporting on the update adds a specific example YouTube gave: tutorial videos that reproduce content already prevalent on the platform rather than offering anything original. Matt Halprin, YouTube's trust and safety lead, framed the problem as a volume effect: the same tool that lets one person make good videos faster also lets them make "lots of videos really quickly that are very similar."
The version that passes: keep the template for structure and replace the substance. A ten-video series on the same topic clears the bar when each video contains a number, a test result, a screen recording, or an outcome that exists nowhere else. It fails when the only variable across videos is the subject noun.
Category two: off-putting or distressing content
This one has nothing to do with AI, and it will catch channels that never used a model. The category covers videos engineered to upset viewers, and the example given is the animal-in-distress-then-rescued format that farms watch time from discomfort. The economics are what made it common: distress holds attention, and attention was paid.
The version that passes: if the emotional peak of your thumbnail and first fifteen seconds is the viewer's discomfort rather than their curiosity, rebuild the opening. Manufactured jeopardy is now a monetisation risk rather than a retention tactic.
Category three: AI personas on sensitive topics
The narrowest and clearest of the three. The policy page states that channels using AI-generated personas presenting themselves as human experts on health, legal, financial or political matters cannot monetise. The July clarification lists finance, legal, healthcare and medical as the sensitive set. Note the two conditions that have to combine: a synthetic persona and an expert claim in a sensitive field. A synthetic narrator reading a recipe is not this. A synthetic doctor explaining a drug interaction is.
The version that passes: keep the synthetic voice or avatar if it helps, drop the implied credential. Attribute the substance to a named human or a cited source on screen, and let the presentation layer be obviously a presentation layer. This is the same distinction Google draws in Search, where the scaled content abuse policy also turns on scale, purpose and value rather than on which tool typed the words.
What suspension actually costs, in re-earned thresholds
The consequences ladder on the policy page has three rungs: limited or no ad earnings on individual videos, suspension from the YouTube Partner Program, and permanent channel termination in exceptional cases. Most channels that get hit land on the middle rung, and the middle rung is where the real cost sits, because YPP is not a switch you flip back on.
Getting back in means meeting the published eligibility thresholds again: 1,000 subscribers plus 4,000 valid public watch hours in the last 12 months, or 1,000 subscribers plus 10 million valid public Shorts views in the last 90 days, along with no active Community Guidelines strikes, 2-Step Verification enabled, and a linked AdSense account. The watch-hour path has a definition worth checking before you count your own: private videos, unlisted content, Shorts views in the Shorts feed, unconverted livestreams and ad campaigns do not count toward the 4,000.
Run that as a rebuild estimate rather than an abstraction. Four thousand valid watch hours is 240,000 minutes. A channel averaging four minutes of watch time per view needs roughly 60,000 qualifying views inside a rolling twelve months to re-clear the bar. Those are our numbers from YouTube's published thresholds, and they will differ for your channel, but the order of magnitude is the point: the penalty is not a missed month of revenue, it is a year of audience-building repeated with no revenue attached.
The two-minute test before you upload
The policy's own language gives a usable test, and it is a substitution test rather than a quality judgement.
Take your script and swap the topic for a different one in the same niche. If the only edits required are proper nouns, the video is the thing the policy describes.
A video that survives the swap contains something that does not transfer: your measurement, your screen, your customer's complaint, the thing that broke when you tried it. That is also the only asset in this whole arrangement that a platform cannot reprice, which is the argument for not letting one payout channel carry the business. YouTube's rules changed by clarification, with no notice period and no appeal window worth planning around, and the same is true of every platform that pays you. The hedge is the same as it has always been: publish where the rules are yours, and move the audience onto a list you own before you need to.
Discussion
Sign in with Google or just a name. No email link, no password to remember.