Brazil: Millions in profits with Gisele Bündchen deepfakes

Brazil is taking action against the increasing use of AI tools to manipulate images and videos of celebrities for online fraud. Last week, police in five Brazilian states executed several arrest warrants and froze assets linked to such online fraud. This was reported by the Brazilian daily newspaper O Globo.

Operation “Modo Selva” (“Jungle Mode”) was directed against a group of suspected fraudsters who used deepfakes to commit fraud worth more than 20 million reais (3.2 million euros). According to the police, the group used artificial intelligence (AI) to produce fake videos of Brazilian model Gisele Bündchen and other celebrities and used Instagram ads to trick victims into buying products, charging only for shipping costs.

In August, Gisele Bündchen denounced the misuse of her image by AI on social networks. At the time, the model warned her followers about sponsored content that was actually fake. “There are several accounts using my picture and my voice manipulated by artificial intelligence to make fake advertisements. Be careful so you don't fall victim to a scam. Report it!” she wrote.

Social media platforms are liable for content

In June, Brazil's Supreme Court ruled that social media platforms can be held directly liable for illegal content posted by their users. A prior court order to delete such posts is not required. Meta, the parent company of Instagram, explained, according to the news agency Reuters, that its guidelines prohibit advertisements “that fraudulently use public figures to deceive people.” Such ads would be removed “upon discovery.” The company also has “special systems in place to recognize celebrity bait.” The company explained that it “invests heavily in trained verification teams, shares tips on how to avoid fraud, and offers tools to report potential offenses.”

According to O Globo, the investigation in Brazil began after a victim of a scam worth just 44.57 reais (7.12 euros) contacted the police. The victim had been deceived by an Instagram ad showing a manipulated video of Bündchen advertising a “free anti-wrinkle set.” “The criminals had created a deepfake video with the image and voice of Gisele Bündchen, making it appear that the model was actually advertising the non-existent product,” the police said.

According to Reuters, most victims lost only small amounts of money, usually less than 100 reais (16 euros), and generally did not report the offenses. “This led to a perverse situation in which the criminals enjoyed a kind of 'statistical immunity'. They knew that most people would not report them and therefore acted without fear on a grand scale,” explained a police spokeswoman in a statement.

(akn)