Mark Cuban: 'AI-Generated Fake Ads Are Using My Face'. Meta's Handling Under Scrutiny

Mark Cuban has accused Meta of knowingly leaving AI-generated ads that impersonate him up on Facebook. Speaking on the Pivot podcast in August 2026, Cuban said ads using a convincing copy of his face and voice were promoting a diabetes treatment. Earlier Reuters reporting on Meta’s internal documents has resurfaced alongside his comments, turning impersonation ads into a controversy once again.

What Cuban is objecting to

Cuban said he has never sold or endorsed the product. Even so, he said he hears every week from people who ordered the $294 item. The harm starts when someone sees an ad featuring a celebrity’s face and voice and believes that person genuinely endorsed the product.

His criticism went beyond whoever made the ads to the platform that carried them. He argued that Meta allows ads like these to run even though it knows the damage they cause. Meta bans ads that use celebrity images to deceive people. The question Cuban is raising is whether that ban actually works in practice.

What the $16 billion in scam ad revenue refers to

In November 2025, Reuters published a report based on a review of internal Meta documents. According to the report, an internal estimate from late 2024 put Meta’s revenue that year from scam ads and ads for banned goods at roughly $16 billion. That is about 10 percent of its total revenue.

Meta pushed back on the report. A company spokesperson argued that the documents show only a slice of its anti-fraud work and therefore misrepresent the company’s approach.

In a separate announcement, Meta said it removed more than 159 million scam ads in 2025. It said 92 percent of those were acted on before any user reported them. The company presented those takedown numbers as evidence that it is working to stop scam ads.

How it plans to block them in Korea

Meta has described technology that screens suspected scam ads by comparing the faces in them against the profile photos of public figures. In September 2025, the company said it would extend that technology to Korea to detect and block celebrity impersonation scams there. The aim is to protect users in Korea from ads that misuse celebrities’ likenesses.

The open question is how quickly reported impersonation ads actually come down. Takedown counts alone say little about whether users are genuinely protected. What is worth watching is how the ads Cuban pointed to are handled, and the status of the blocking technology’s rollout in Korea.