OpenAI says Chinese accounts tried to turn Americans against data centres (www.engadget.com)

OpenAI has released a report indicating that Chinese-linked fake accounts utilized ChatGPT to orchestrate a campaign aimed at influencing American perceptions about AI data centers. The report identifies two clusters of users; one group, dubbed the "Data Center Bandwagon," generated English-language content criticizing the electricity demand from AI data centers and its impacts on consumer costs. Posing as American citizens, these accounts engaged on social media, using ChatGPT to create tailored talking points and visuals, while leveraging legitimate news links to lend authenticity to their claims. The significance of this revelation lies in the covert nature of these influence operations, which sought to intervene in the U.S. discourse on AI without revealing their true affiliations. The second cluster focused on critiquing U.S. tariffs and tech policies, particularly targeting Taiwanese audiences, showcasing a sophisticated use of language and imagery across multiple platforms. Although OpenAI noted that the campaigns did not manage to sway public opinion significantly, the incidents highlight ongoing vulnerabilities in the online discourse surrounding AI, raising concerns about the manipulation of societal conversations using advanced AI tools.