AI-made supplement adverts turn racist shock tactics into a production line

AI-generated musical cartoons are helping supplement seller Resilia produce thousands of online adverts built around racist stereotypes and sensational health claims, CNN reports. The scale shows how cheap generative tools can turn attention-grabbing nonsense into a high-volume advertising strategy, while leaving platforms to catch up after the fact.

Watch Desk analysis

What happened

CNN reports that Resilia launched more than 8,000 ads on Google over the past year and a half, including about 7,000 on YouTube. The ads use AI-generated animation and music to spin long, lurid stories before revealing products such as garlic capsules. Some feature antisemitic tropes, while others use racial stereotypes and claims about weight, health or longevity.

Marketing researchers quoted by CNN say generative AI makes it cheap to produce and test many variations, rather than betting on a single polished advert. Google told CNN it removed ads flagged for policy violations, including clickbait and deceptive health claims. One ad Google said it had removed appeared on YouTube again the following day, and the company said it was investigating. Meta removed a number of Resilia adverts after being alerted to them; CNN reports the count on the platform fell from around 1,500 in mid-September to fewer than 500. Resilia CEO Johnathan Sack did not respond to CNN’s requests for comment.

Why it matters

The AI connection is not cosmetic: automated production lets advertisers generate and test a volume of strange, provocative creative that would be expensive to make by hand. The ads can attract attention through outrage before they sell anything, and the health claims matter because the products are supplements people may ingest. CNN notes that Resilia says its products have not been evaluated by the US Food and Drug Administration.

The figures and examples come from CNN’s reporting, not a published audit of ad performance or a finding that every advert violated platform rules. But the reported reappearance of an ad after removal points to a familiar enforcement problem: taking down one version may not be enough when variants are cheap to produce.

Our read

This is AI doing something less glamorous than composing a sonnet and rather more consequential than making a funny picture: multiplying the supply of persuasive, potentially misleading advertising. Platforms should explain how quickly they can detect repeat variants and what happens when an advert returns. For consumers, a polished cartoon is still an advert, however much plot and backing music it brings along.

What to watch

• Whether Google and Meta disclose how they identify and remove related ad variants.
• Whether Resilia responds to CNN’s account or changes its advertising practices.
• Whether regulators or platforms publish clearer evidence about health claims and AI-generated adverts.

Discussion spark: Should platforms be required to catch and block new AI-generated versions of an advert after removing one, or is that an unrealistic standard for moderation?

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.