Brazilian Police crack down on Deepfake fraud ring: Four arrested in $3.9 million celebrity scam

Police in Brazil have delivered a significant blow to a sophisticated online fraud network, arresting four suspects and seizing assets as part of a probe into a multi-million-dollar deepfake scam. The criminals allegedly used AI-generated videos of supermodel Gisele Bündchen and other celebrities in fraudulent Instagram advertisements, tricking thousands of users out of over 20 million reais (approximately $3.9 million).

This operation, led by the cybercrime unit in the southern state of Rio Grande do Sul, marks one of the country's first major efforts to combat the criminal exploitation of AI technology in celebrity-related scams.

The AI-powered scheme

The complex scheme centered on creating convincing deepfake videos that appeared to show famous figures like Bündchen endorsing bogus products and nonexistent giveaways. The scam advertised these fraudulent offers via paid ads on Instagram, leveraging the celebrities' trusted public images to deceive consumers.

The police investigation, which began in August 2024 after a victim reported being duped, identified two primary tactics:

• False endorsements: One victim was misled by an Instagram ad featuring an altered video of Gisele Bündchen promoting a skincare product.

• Fake giveaways: Another tactic involved using the supermodel's likeness to promise a suitcase giveaway, with victims paying small, non-existent "shipping fees" for items that were never delivered.

Investigators confirmed the criminal group was also involved in a series of other scams, including deepfakes of additional celebrities and fraudulent betting platforms, illustrating the network's broad reach.

Mass Fraud fueled by 'Statistical Immunity'

While authorities successfully identified more than R$20 million in suspicious funds generated by the ring, investigators noted that most individual victims lost a relatively small amount, often under 100 reais ($19). This low individual loss played directly into the criminals' strategy.

“That created a perverse situation in which the criminals enjoyed a kind of 'statistical immunity,'” stated Isadora Galian from the cybercrime unit. “They knew most people would not report them, so they operated at scale without fear.”

The four suspects were arrested this week, and authorities have taken action to freeze assets across five different Brazilian states. The suspects face charges including online fraud and money laundering.

Social media platforms under scrutiny

The case highlights the growing global challenge social media companies face in policing AI-driven deceptive content.

Meta, the owner of Instagram, issued a statement confirming that its policies explicitly prohibit "ads that deceptively use public figures to try to scam people," and that it removes such content "when detected." The company added that it employs "specialised systems to detect celeb-bait" scams.

Supermodel Gisele Bündchen had previously alerted her followers to the circulation of fake videos online using her image for scams, urging her fans to remain vigilant against such fraudulent activity.