Influencer Deepfake Ads Are a Likeness and Trust Problem, Not Just Spam

An influencer deepfake ad is a paid social post built from AI-altered or AI-generated imagery of a real creator’s face, run to promote a product that creator never endorsed. Per The Guardian’s September 12, 2026 report, Emily Schuman’s fake GLP-1 ad for telehealth brand Gala was built from an old car selfie she had posted years earlier, altered to show her holding the product.

The same reporting documents a second version of Schuman with darker hair applying foundation for Meroda Cosmetics, plus TikTok posts using generated images of her promoting Superpower’s blood-testing service. Three different fake campaigns, one real face, none of them authorized.

Emily Schuman, founder of Cupcakes and Cashmere with more than 500,000 Instagram followers as of The Guardian’s September 12, 2026 report, discovered an Instagram ad using her AI-altered likeness to promote Gala, a telehealth brand she had never worked with. Followers messaged her directly about the off-brand sponsorship.

“It’s so violating.”

Emily Schuman, founder of Cupcakes and Cashmere, describing the fakes to The Guardian as looking “just like me, but the upside down version.” Neither Gala nor Meroda Cosmetics responded to The Guardian’s requests for comment on the ads using Schuman’s likeness.

Slower than the policy text suggests. Meta and TikTok both prohibit using AI or digital manipulation to replicate a person’s identity without consent, but Schuman and her manager needed weeks of sustained pressure before Meta removed the fake Gala ad, and the Meroda foundation ad was still circulating on Instagram despite repeated takedown reports, per The Guardian.

Meta has added advertiser verification and AI detection tools aimed at what it calls celeb-bait, and did not respond to The Guardian’s request for comment on Schuman’s case. Experts quoted in the same report say those tools do not go far enough. For a brand-safety team, the operating fact is simple: platform policy bans the behavior, but removal speed, not policy language, is what a follower actually experiences.

A wider pattern. The Guardian’s reporting also names beauty influencer Arielle Lorre, deepfaked for skincare ads; fashion creator Jessi Caparella, who found altered images on Pinterest pointing to affiliate clothing links; and deepfaked OnlyFans videos used to route fans toward fraudulent accounts. Separately, fashion influencer Molly Tranchin (FashionVeggie) sued underwear brand Eby over an alleged deepfake edit of contracted creative, later dropping the suit for jurisdiction reasons with plans to refile.

Superpower CEO Max Marchione confirmed the TikTok post using Schuman’s likeness was not a real company ad. “Whilst scamming isn’t new, AI has made this much worse,” he told The Guardian, after Superpower reported the scam post itself. Alice Marwick, director of research at the Data & Society Research Institute, frames the business stakes plainly: for an influencer, “their brand is their business,” so a hit to the likeness is a hit to the commercial life around it.

A brand that finds an unauthorized AI likeness of a creator running in its own paid social, or hears about one from confused followers, should treat the first 48 hours as a brand-safety incident with a clock, not a customer-service ticket.

1. Assume scam or scrape first: Superpower's response is the clean model: confirm the ad was not bought by the brand, say so publicly, and report it to the platform the same day, rather than leaving the named creator to field the confusion alone.
2. Audit paid social before blaming the creator: Search for unauthorized creator lookalikes running against the brand's own products before assuming an existing partner went off-brand. The likeness may not belong to anyone the brand has ever worked with.
3. Report and track removal on a clock: File the platform takedown request immediately and track it. Weeks-long removal, as Schuman experienced with both the Gala and Meroda ads, is already a trust loss even after a report is filed.
4. Publish a same-day denial: If a fake ad using a real creator's face surfaces under a brand's name, a same-day public statement disclaiming it limits the window where followers assume the brand is complicit.

A brand cannot answer whether an unauthorized AI likeness of a partnered creator is running against its own name by waiting for a follower screenshot to go viral. Storika tracks creator identity, disclosure status, and campaign activity across a database of 7M+ creator profiles and 80M+ analyzed posts, so a brand can confirm what a named creator actually posted or authorized before an unauthorized version spreads further.

Pair this with AI likeness rights in creator marketing for the consent and provenance side of the same problem, and Huawei’s no-disclosure creator brief for the undisclosed-paid-ask lane. Likeness theft is “someone stole my face into an ad.” A no-disclosure brief is “someone asked me to hide the paid relationship.” Both erode follower trust, and both need separate playbooks.

Pair this guide with AI Likeness Rights in Creator Marketing, A Huawei-Branded Brief Told Creators to Skip Paid Tags, and Influencer Fraud Detection for the adjacent brand-safety, disclosure, and audience- quality risks that show up alongside likeness theft.