What Exactly Did Google Add?
One sentence, placed as a note under the general unacceptable-business-practices rules. The full text: “Note: Ads that impersonate public figures using AI aren’t allowed, even if they disclose that they were generated or edited using AI” (Google Ads policy). Google’s November 2026 update entry describes the change as a clarification meant to help advertisers understand the policy, and says it “does not change the enforcement of the policy” (Google Ads policy update). Search Engine Roundtable first reported the change on November 12 (Search Engine Roundtable).
The sentence answers one specific question: whether disclosing AI use makes an impersonation acceptable. Google’s answer is no: if the ad pretends to be, or implies a connection to, a public figure, it breaks the rule whether or not it admits to using AI.
Who Counts as a Public Figure Under Google’s Rule?
More people than most executives assume. The policy’s public-figure section defines them as “people who are famous or well known” and lists “government officials, entertainers, athletes, online influencers and prominent business people like founders or CEOs of big companies” (Google Ads policy). One of its examples is AI-generated video of popular YouTubers used to suggest they endorse a get-rich-quick scheme.
For a PR team, the line about founders and CEOs matters most. Cloned executive voices are already used in fraud, as our explainer on how a cloned voice got a bank to wire €95 million describes. Google’s text does not set a fame threshold, so whether a mid-size company’s CEO qualifies is, in practice, Google’s call.
How Do Google, Meta and TikTok Compare?
All three ban fake endorsements by public figures; they differ on whether an AI label helps. We read each platform’s current ad-policy text on November 12, 2026 and set the operative language side by side.
| Platform | Rule on public-figure likeness | Does an AI label help? | Stated penalty | Page dated |
|---|---|---|---|---|
| Google Ads | Bans pretending to be a public figure or implying a connection “when you don’t” have one | No: banned “even if they disclose” AI use | Account suspension without warning; no return | Note added Nov 11, 2026 |
| Meta (Facebook, Instagram) | Bans celebrity-bait: “falsely using the name, likeness, or imagery of a public figure” with misleading tactics | Not addressed; the page does not mention AI | Not stated on the policy page | Oct 29, 2026 version |
| TikTok Ads | Bans content that “uses or promotes the misuse of the likeness” of a public figure without permission | Labels are required for AI content, but they do not cure a likeness violation | Undisclosed AI content “will be rejected or restricted” | Last updated April 2026 |
Meta’s rule sits in its Prohibited Commercial Practices standard, which it says now covers ads and organic posts with one rulebook (Meta Transparency Center; Meta Prohibited Commercial Practices). TikTok separately requires an AI label or disclaimer on ads with significantly edited or AI-generated media, including “Making the primary subject say something they didn’t actually say, using AI voice-cloning” (TikTok Ads policy). Google is the only one of the three whose text now answers the label question directly.
How Hard Does Google Enforce It?
It is classed as an egregious violation, with immediate suspension. The policy says “This suspension occurs upon detection and without prior warning” and “You will not be allowed to advertise with Google Ads again” (Google Ads policy). Google’s 2024 Ads Safety Report, published April 16, 2025, said a team of more than 100 experts worked on public-figure scams and that Google permanently suspended “more than 700,000 offending advertiser accounts,” which “led to a 90% drop in reports of this kind of scam ad” (Google blog). Those are Google’s own figures; we have no independent count.
Legitimate advertisers are affected too. Google’s best-practice text asks advertisers to hold provable consent before using a public figure’s name, image, video or voice to suggest an endorsement. On appeal, it accepts contracts, written agreements, credible news coverage of the relationship and public statements from the figure.
What Should PR Teams and Executives Do Now?
Keep proof of consent for any real endorsement, and report fakes through the public-figure route. Our recommendations, not platform rules: if your executive appears in a partner’s ads, put the permission in writing and include AI-edited versions explicitly, since Google may ask for proof on review. If a fake appears, Google’s policy page links a public-figure report form; our step-by-step guide to reporting a scam ad impersonating your company covers Google, Meta, Microsoft, X, TikTok and LinkedIn, and our explainer on whether a platform can be held liable covers what happens when reports fail.
Methodology and Limitations
Four limits apply. First, we compared published policy text only, not enforcement in practice. Second, Meta’s summary page and its full standard are dated differently (September 9 and October 29, 2026), and neither mentions AI, so our reading of Meta’s position on labels is limited to what the text omits. Third, Google says the note does not change enforcement, so it should not be read as a new ban. Fourth, Google’s enforcement figures are self-reported and a year old. The next thing to watch is Google’s 2026 Ads Safety Report, which on past timing would arrive in spring 2027. This is not legal advice.
Sources and Methodology
This explainer is built from Google’s Unacceptable Business Practices policy and its November 2026 update entry, Meta’s Prohibited Commercial Practices pages, TikTok’s misleading and false content ad policy, and Google’s 2024 Ads Safety Report, all read November 12, 2026. The comparison table is our reading of each text. Recommendations are ours, not platform rules. We did not test ad submissions.
- platform documentation Google Ads, Unacceptable business practices. The AI note, public-figure definition, examples, enforcement and appeal text. Read November 12, 2026.
- platform documentation Google Ads, Updates to Unacceptable Business Practices (November 2026) (November 11, 2026). Date and “does not change the enforcement.” Read November 12, 2026.
- platform documentation Meta Transparency Center, Prohibited Commercial Practices (October 29, 2026 version). Celebrity-bait definition. Read November 12, 2026.
- platform documentation Meta Transparency Center, Our Prohibited Commercial Practices policy (updated September 9, 2026). One rulebook for ads and organic content. Read November 12, 2026.
- platform documentation TikTok, Misleading and false content (last updated April 2026). Likeness and AI-label rules. Read November 12, 2026.
- vendor Google, Our 2024 Ads Safety Report (April 16, 2025). 700,000 suspensions, 90% drop in reports. Read November 12, 2026.
- trade press Search Engine Roundtable, Google Ads Disallows Impersonate Public Figures With AI Even If Disclosed (November 12, 2026). News context. Read November 12, 2026.