Public figure impersonation is the use of a famous person's name, image, voice or synthetic likeness in an advertisement to make people believe that person is behind, endorses or is connected to an offer when they are not, so that viewers hand over money or personal information. The phrase is above all a policy label. Google wrote it into the unacceptable business practices section of its Misrepresentation policy in 2024, and it sits on the short list of offences that end an advertising account on detection, without warning. It exists because borrowed fame is the cheapest trust signal a scammer can buy, and generative artificial intelligence (AI) has made it almost free to forge.

What the rule covers

Google's rule turns on three elements. The ad must involve a public figure, either by pretending to be that person or by "making it seem like you have a connection to one when you don't", according to the Advertising Policies Help Center. The connection must be false. And the aim must be to get people to part with money or personal information.

Public figures are defined as "people who are famous or well known": government officials, entertainers, athletes, online influencers and prominent business people such as founders or chief executives of large companies. A famous singer's photograph used without permission to suggest they are a happy client of an investment service is a violation. So is a false claim that a professional athlete sponsors a weight-loss product, and so are AI-generated videos of popular YouTubers made to look like endorsements of a "get rich quick" scheme.

The third element does a lot of work. A satirical ad featuring a politician is not caught simply because a recognisable face appears. A fake interview in which a television presenter "reveals" a trading platform, linked to a page that takes deposits, is the archetype. Google's stated best practice is documented consent that the advertiser can produce on request.

The AI clarification. On November 11, 2026, Google posted "Updates to Unacceptable Business Practices (November 2026)". The policy page now says "ads that impersonate public figures using AI aren't allowed, even if they disclose that they were generated or edited using AI". Google said the update does not change enforcement, and the notice did not address licensed likenesses or parody. An AI disclosure label describes how an ad was made; it does not cure a false claim of endorsement.

How enforcement works

Severity defines the category at Google. Most policy breaches lead to a disapproved ad and, if repeated, a three-strike ladder with a seven-day warning before suspension, according to Google's documentation. Public figure impersonation skips that ladder. As an egregious violation, one Google describes as unlawful or posing significant harm to users, it brings suspension "upon detection and without prior warning", and the advertiser may not return. Related accounts, and new ones the same advertiser opens, can be suspended too.

Detection runs in layers:

  1. Identity checks. Advertiser verification ties an account to a legal entity or person, and Google clarified in November 2025 that false verification details lead to suspension without warning.
  2. Automated review of creative, landing pages and account signals.
  3. Face matching. Meta compares faces in suspect ads with enrolled public figures' profile pictures, a system it began testing in October 2024 and which covered more than 500,000 people when it sued scam advertisers in Brazil, China and Vietnam on February 26, 2026. YouTube's likeness detection applies similar matching to uploaded videos rather than ads.
  4. Reports from users and from the people impersonated.

Recourse is narrow. Accounts "are only reinstated in compelling circumstances, such as in the case of a mistake", Google says, and on July 21, 2026 it ended in-account appeals for decisions more than six months old. Scammers respond with cloaking, showing reviewers a harmless page, and with rented accounts whose clean histories pass trust checks.

Origin and evolution

Fake celebrity endorsements predate the policy. In April 2018 consumer journalist Martin Lewis sued Facebook for defamation over more than 1,000 scam ads carrying his face, settling on January 23, 2019 for a £3 million donation to Citizens Advice and a UK scam ad reporting button. The Amsterdam District Court ordered Facebook on November 12, 2019 to act against fake bitcoin ads using John de Mol, and Australia's competition regulator, the ACCC, sued Meta on March 18, 2022.

On February 15, 2024, Google announced that unacceptable business practices would explicitly cover impersonating public figures, brands or organisations to entice people "to part with money or information". PPC Land reported that the update meant immediate and permanent suspension for impersonators. Enforcement began in March 2024, and in April in France, moving the offence from the disapproval tier to the account-ending one.

The US Federal Trade Commission (FTC) finalised its rule on impersonation of government and businesses the same day, and it took effect on April 1, 2024. The FTC also proposed extending it to impersonation of individuals, which would reach celebrities, and to firms providing the "means and instrumentalities" of impersonation. It dropped the second idea on December 26, 2024. The first went to an informal hearing in January 2025, and the agency is "still evaluating its options", according to the Federal Register.

YouTube launched likeness detection for Partner Program creators in October 2025, extended it to government officials, journalists and candidates on March 10, 2026 and opened it to every creator aged 18 and over in May 2026.

Why it matters for the marketing community

Scale comes first. In its 2024 Ads Safety Report, released in April 2025, Google said more than 100 experts worked on the problem, that it permanently suspended more than 700,000 advertiser accounts for public figure impersonation, and that reports of the scam fell by 90%. Google removed 146.9 million ads under the Misrepresentation policy that year and 421.5 million in 2025, when it gave no impersonation-specific count.

Money follows. Internal Meta documents reported by Reuters in November 2025 projected that about 10% of 2024 revenue, roughly $16 billion, came from scam and banned-goods ads, and PPC Land covered the penalty bids charged to suspected fraudsters. Meta called the figure "rough and overly-inclusive". Impersonation ads bid for the same inventory as genuine campaigns and erode trust in celebrity endorsement.

Legitimate advertisers carry risk too. Because the category ends accounts without warning, a misjudged creative costs more than almost any other policy mistake, so agencies using licensed talent or AI production now keep consent records as carefully as media plans.

Limitations, criticisms and disputes

Definitions. Google sets no threshold for fame. A regional financial commentator may fall on either side, decided by classifiers no advertiser can inspect.

False positives. Licensed talent, accurate quotes and commentary about officials can trip the same detectors. Google claimed in November 2025 it had cut incorrect suspensions by 80% without publishing its method or absolute numbers.

Self-reported results. The 700,000 suspensions, the 90% drop and Meta's roster size are unaudited company figures, and falling reports can mean users stopped reporting.

Coverage gaps. Face matching protects only enrolled people with usable reference images, and runs only after other classifiers flag an ad.

Liability. The core dispute is whether a platform that auctions and optimises an impersonation ad is a neutral host. On September 16, 2026 the Frankfurt Regional Court held that Meta's auction and ranking gave it enough control to lose its hosting exemption under the Digital Services Act, with fines of up to €250,000 per repeat violation. The ruling is not final. In the US the question runs through Section 230 in Andrew Forrest's suit against Meta.

  • Celeb-bait is Meta's name for the scam tactic. Public figure impersonation is the wider policy category, covering officials, executives and influencers as well as celebrities.
  • Right of publicity is the legal right to control commercial use of one's identity. An unlicensed but honest ad can breach it without being a scam; impersonation requires deception aimed at money or data.
  • Deepfake names a synthetic media technique. Many impersonation ads use plain doctored photos, and most deepfakes are not ads.
  • Brand and government impersonation sit in the same Google sub-policy and the FTC's 2024 rule, but concern organisations. In US federal rulemaking they are covered where individuals are not.

Recent developments

Ofcom proposed fraudulent advertising codes under the UK Online Safety Act on July 10, 2026, naming public figure impersonation and deepfakes as risk signals. The NO FAKES Act, which would create a federal right against unauthorised digital replicas, cleared the Senate Judiciary Committee on June 18, 2026, before Senator Ted Cruz blocked unanimous consent on September 30. Denmark's bill L 49, tabled on October 8, would require publishers to prove consent for realistic digital imitations from January 1, 2027.

The FTC published an advance notice of proposed rulemaking on October 1, 2026, asking whether platforms that "profit from optimizing online ads for third parties, regardless of whether the third parties are legitimate" should carry duties under the impersonation rule. Options include reviewing ads before posting, denying optimisation tools such as Meta's Advantage+ to impersonators, and verifying advertisers first. The notice covers government and business impersonation only; comments close on November 30, 2026. Reported imposter losses reached nearly $3.5 billion in 2025, according to the notice.

Disclosure rules now overlap with impersonation rules. Article 50 of the EU AI Act has required deepfakes to be disclosed since August 2, 2026, though a Resemble AI scan found only 2 of 3,379 AI-flagged Meta ad files disclosed AI use. Google's November 11 clarification settles one question that labelling raised: a disclosed fake endorsement is still a fake endorsement.

Timeline

  • April 2018 - Martin Lewis sues Facebook over more than 1,000 scam ads using his name and face.
  • January 23, 2019 - Lewis settles; Facebook pledges £3 million to Citizens Advice and a UK scam ad reporting button.
  • November 12, 2019 - Amsterdam District Court orders Facebook to act against fake bitcoin ads using John de Mol.
  • March 18, 2022 - ACCC sues Meta over crypto scam ads using Dick Smith, David Koch and Mike Baird.
  • February 15, 2024 - Google announces public figure, brand and organisation impersonation as unacceptable business practices; the FTC finalises its government and business impersonation rule and proposes covering individuals.
  • March 2024 - Google begins immediate suspension for public figure impersonation outside France; April 2024 in France.
  • April 1, 2024 - FTC impersonation rule takes effect.
  • December 26, 2024 - FTC drops the proposed means-and-instrumentalities provision.
  • January 2025 - FTC holds an informal hearing on extending the rule to individuals.
  • October 21, 2024 - Meta begins testing facial recognition against celeb-bait ads.
  • April 2025 - Google's 2024 Ads Safety Report records more than 700,000 accounts suspended for public figure impersonation.
  • October 2025 - YouTube launches likeness detection for Partner Program creators.
  • February 26, 2026 - Meta sues scam advertisers in Brazil, China and Vietnam; its protection roster exceeds 500,000 public figures.
  • March 10, 2026 - YouTube extends likeness detection to government officials, journalists and candidates.
  • May 2026 - YouTube opens likeness detection to all creators aged 18 and over.
  • June 18, 2026 - Senate Judiciary Committee advances the NO FAKES Act.
  • July 10, 2026 - Ofcom proposes fraudulent advertising codes.
  • August 2, 2026 - EU AI Act Article 50 transparency duties apply.
  • September 16, 2026 - Frankfurt Regional Court rules against Meta over fake Finanzfluss ads.
  • September 30, 2026 - Senator Ted Cruz blocks unanimous consent for the NO FAKES Act.
  • October 1, 2026 - FTC publishes an advance notice on platforms' role in impersonation scams; comments due November 30.
  • October 8, 2026 - Denmark tables deepfake bill L 49.
  • November 11, 2026 - Google clarifies that AI disclosure does not excuse public figure impersonation.

Summary

Who. Scam operators who borrow the identities of entertainers, athletes, influencers, executives and officials; the platforms that police them, led by Google, Meta and YouTube; and regulators including the FTC, Ofcom and European courts.

What. Advertising that pretends to be, or falsely implies a connection to, a well-known person in order to obtain money or personal information. At Google it is an egregious breach of the Misrepresentation policy, punished by permanent suspension without warning, and AI disclosure does not excuse it.

When. Litigation over fake celebrity endorsements began in 2018. Google created the explicit category on February 15, 2024 and enforced it from March 2024, and clarified its application to AI-generated content on November 11, 2026.

Where. Search, video and social ad platforms worldwide, with the sharpest legal tests in the UK, the Netherlands, Australia, Germany, Denmark and the US.

Why. A familiar face lowers suspicion faster than any copy. Self-serve ad systems make that trust cheap to rent and generative AI makes the likeness cheap to forge, so platforms have moved the offence to the top of their penalty scale while regulators debate how much of the burden the platforms themselves should carry.