The cover of the April 15 issue of Die Aktuelle proclaims the German tabloid contains “the first interview” with Formula 1 champion Michael Schumacher.
Getting the famed race car driver, who has rarely been seen in public since a serious skiing accident in 2013, to answer questions would have been a huge scoop. Yet those who read the story would have learned that Schumacher’s answers weren’t from the driver at all, but rather from an A.I. chatbot trained to emulate the race car driver.
“It sounded deceptively real!” the tabloid wrote, advertising the interview that featured the ersatz Schumacher answering questions about his time in hospital and his recovery.
Schumacher’s family is not amused with the tabloid’s experiment, confirming to Reuters that they are planning legal action.
The Schumachers have battled with Die Aktuelle before. In 2015, a court dismissed a lawsuit by the family after the tabloid released an issue featuring Corinna Schumacher, Michael Schumacher’s wife, under the headline “A New Love Makes Her Happy.” The publication later revealed the story was actually about the Schumachers’ daughter.
Media reports suggest that Die Aktuelle used a chatbot from A.I. service Character.AI, and screenshots of the tabloid’s conversation with “Schumacher” appear to be from its platform. Character.AI allows users to create and train their own chatbots, often emulating real-world celebrities, historical figures, and fictional characters. (Fortune has reached out to Character.AI.)
Die Aktuelle, in its un-bylined piece, is cagey about how the Schumacher bot was developed, suggesting that “someone from the family, carers, or employers,” or even Schumacher “from the hospital bed” could have trained it.
A.I. programs could unleash a new wave of misinformation and “deepfakes,” as increasingly sophisticated models generate more realistic creations. In late March, an A.I.-generated image of Pope Francis wearing a white puffer jacket went viral as users were fooled by the Pope’s (A.I.-generated) fashion sense.