Google Ads Reportedly Says an AI Label Won't Excuse Impersonating a Public Figure

Google permanently suspended more than 700,000 advertiser accounts for impersonating public figures in 2024, according to a published report, and a new help-center notice rules out an AI label as a defense.

An advertiser that turned on Google’s AI label and read it as cover now holds an account exposed to suspension from a single ad, with no chance to pull the ad first. Enforcement did not change, the report said. What changed is the reading: a label tells viewers how an ad was made, while the rule asks whether the ad puts a real person behind a product falsely.

What did Google say?

The notice calls itself a clarification that “does not change the enforcement of the policy,” the report said. The policy page already put impersonation among the offenses Google calls egregious: the account is shut the moment Google detects the ad, the advertiser cannot return, and reinstatement is rare.

Google counts a public figure as anyone famous or well known, from officials and athletes to influencers and company founders. Its examples already included AI-generated videos of popular YouTubers made to suggest they back a get-rich-quick scheme, so the open question was the label alone.

Why did the label raise the question?

Google added a “How this ad was made” panel in July, the report said, along with AI label settings across its ad products. An advertiser who switched the label on was following Google’s own disclosure tools, and the notice says that does not protect the ad.

Resemble AI published a study in October of Meta ads that its own detection model flagged as AI-made: only 2 of 3,379 files disclosed AI use. The study covers Meta’s platforms, not Google’s, and the report read it as a sign of how rarely synthetic ads carry any label.

What does the notice leave open?

Licensed likenesses. Google did not say whether an AI-made ad featuring a public figure who agreed to it, such as a brand’s contracted ambassador, is treated differently. The policy page advises advertisers to hold documented permission before using a public figure’s name, image, video or voice, but the notice does not link that advice to AI.

Google also did not publish revised policy wording or say what prompted the clarification, the report said.

Does a brand that licenses a celebrity and produces AI variations of the endorsement hold written permission from that person for each ad, before one flag ends the account?