On Instagram, Sartori shared an audio in which the former president presumably sings a popular funk song. However, the voice is totally generated by artificial intelligence.
The journalist warns about the simulation because it can also imitate the person’s face, making it more difficult for untrained people to detect the sham.
“This voice does not belong to a Dilma’s imitator. It was created 100 percent with artificial intelligence. It is just a computer reading a text. In a dark basement, they may now be producing fake contents to be used during the election. The vaccine against misinformation is education,” Sartori wrote.
Creating and sharing fake contents this way is much more complex and refined than what usually circulates on the Internet, he noted.
The journalist pointed out that modality is known as deep fake, and its use poses a threat to Brazilian authorities in the face of the general election on October 2022.
That attack modality could be used during the presidential campaign, a process in which, according to recent polls, the pre-candidate from the Workers’ Party, Luiz Inacio Lula da Silva, is the front runner.