Bots and automated responses produced by ChatGPT circulate freely on X after Elon Musk dismantled the content moderation team and the verification tools used by journalists and researchers, following his acquisition of the platform in October 2022.
Musk's administration marked 22 months three days before the platform was blocked in Brazil in the early hours of Saturday (31). One of the billionaire’s promises during the acquisition process was to put an end to the bot problem, which, in his view, was plaguing the old Twitter.
Bots that reply to public authorities, seemingly using artificial intelligence (AI), are a clear example of the current crisis. Verified profiles are prioritized by the platform's algorithm and appear at the top of response lists to posts by heads of state.
The report identified 20 accounts making thousands of posts in various languages. In some of these, phrases typical of ChatGPT responses appeared, such as "up to the cutoff date of my knowledge" and "as an AI language model." One of these profiles, that of Israeli Tomer Rozenberg, in a single day spoke in French with Canadian Prime Minister Justin Trudeau, commented in Portuguese on the actions of Brazilian President Luiz Inácio Lula da Silva (PT), and conversed in German with WHO (World Health Organization) Director-General Tedros Adhanom. Rozenberg told Folha that he uses ChatGPT to reply to authorities worldwide. With this strategy, the self-help writer gained 20 million views and 3,700 followers in three months. He hopes to use this reach to promote his work.