Google says AI ads impersonating public figures stay banned even when disclosed

Google published a policy notice titled "Updates to Unacceptable Business Practices (November 2026)" on 11 November stating that ads using AI to impersonate public figures are prohibited under its Unacceptable Business Practices policy. The operative text says such ads "are not allowed" and that the prohibition applies "even if they disclose that they were generated or edited using AI," adding that the notice "does not change the enforcement of the policy."

Google framed this as a clarification rather than a new rule. The prohibition on public figure impersonation has been enforced since March 2024, and April 2024 in France, originally under the Misrepresentation policy. No separate effective date or new enforcement date was set. Violations remain classed as egregious, meaning accounts are suspended on detection without warning. Google did not publish the full revised policy wording or a comparison with the previous version, and the notice does not address AI likenesses used with a public figure's consent. Google's own Unacceptable business practices help page carries the same sentence and gives AI-generated videos of popular YouTubers used to imply endorsement of a get-rich-quick scheme as an example of a prohibited ad.

Why it matters for digital humans: This settles a question that AI-disclosure advocates had left open — whether a visible "AI-generated" label cures an unauthorised likeness. On Google's ad platform it does not, which puts consent, not disclosure, at the centre of compliance. For agencies and avatar vendors running paid campaigns, that means a signed likeness release is the thing that keeps an account alive; a watermark or label is not a substitute. The gap worth noting is the one Google did not address: consented AI likenesses of public figures, which is exactly the use case the avatar industry sells.