OpenAI has identified and banned a cluster of ChatGPT accounts originating from China that were used in a covert influence operation targeting the US. The operators, likely part of a Chinese tech company's social media team working for government clients, used VPNs to access OpenAI's platform and generated content in both English and Chinese.

The banned accounts primarily targeted US audiences by creating social media posts and images claiming that data centers and AI applications were driving up electricity demand and costs for ordinary Americans. They used hashtags like #capacityauction and #datacentersuccess, and manipulated images to support their narrative. The operation also targeted overseas Chinese, attempting to harass individuals like dissident Li Ying.

A notable tactic involved posing as US-based Chinese immigrants and workers to encourage influencers to criticize US policies. The operation also leveraged ChatGPT for automating tasks, managing social media interactions, and processing data. Their reports revealed a sophisticated understanding of platform detection systems, focusing on building credible personas and expanding audience reach through organic engagement and advertising, while also planning for account safety and avoiding detection of coordinated activity.