OpenAI Report Says Chinese Official’s ChatGPT Use Exposed Alleged Global Intimidation Campaign
A Chinese law enforcement official’s use of ChatGPT inadvertently exposed what OpenAI describes as a wide-ranging campaign aimed at intimidating Chinese dissidents living abroad, including operations that allegedly impersonated US government officials and used forged legal documents.
In a newly released report, OpenAI said one user — believed to be connected to Chinese law enforcement — used ChatGPT as a kind of operational journal, documenting strategies tied to a broader transnational repression effort.
According to OpenAI investigators, the material described coordinated tactics designed to silence critics of the Chinese Communist Party (CCP) overseas.
Alleged impersonation and forged documents
In one instance detailed in the report, operators allegedly posed as US immigration officials to warn a Chinese dissident living in the United States that their public criticism had violated American law. The aim, according to the ChatGPT entries, was to intimidate the individual into silence.
In another example, the user described fabricating documents that appeared to come from a US county court in an effort to pressure social media platforms into removing a dissident’s account.
OpenAI said it was able to cross-reference descriptions written in ChatGPT with real-world online activity. One documented effort involved spreading false rumors that a Chinese dissident had died. The user reportedly described creating a fake obituary and fabricated images of a gravestone. Similar claims about the dissident’s death later appeared online in 2023, including in Chinese-language reporting by Voice of America.
“Industrialized” information warfare
Ben Nimmo, principal investigator at OpenAI, characterized the operation as a modern example of state-linked transnational repression.
“This is what Chinese modern transnational repression looks like,” Nimmo said ahead of the report’s publication. “It’s industrialized. It’s about targeting critics everywhere, using multiple tactics at once.”
According to OpenAI, the network appeared to involve hundreds of operators managing thousands of fake social media accounts across platforms. While ChatGPT served as a planning and documentation tool, much of the propaganda content was generated elsewhere and then disseminated through coordinated accounts and websites.
OpenAI said it banned the account after uncovering the activity.
CNN has sought comment from the Chinese Embassy in Washington, DC, regarding the allegations.
Attempt to target Japanese leadership
The report also described an effort to influence public perception in Japan. The ChatGPT user allegedly asked the AI to develop a strategy to discredit incoming Japanese Prime Minister Sanae Takaichi, including amplifying online anger over US tariffs on Japanese goods.
According to OpenAI, ChatGPT declined to generate the requested plan. However, in late October — as Takaichi assumed office — hashtags criticizing her and referencing US tariffs surfaced on a popular online forum used by Japanese graphic artists.
OpenAI said the overlap suggests coordinated attempts to shape digital narratives beyond China’s borders.
AI competition and geopolitical stakes
The revelations come amid intensifying strategic competition between Washington and Beijing over artificial intelligence leadership.
The US and China are racing to dominate advanced AI systems, with implications for military applications, economic power, and global influence operations.
The Pentagon is currently engaged in a dispute with Anthropic, another major AI firm. Defense Secretary Pete Hegseth has reportedly given Anthropic CEO Dario Amodei a deadline to adjust safeguards on its AI model or risk losing a Defense Department contract.
Michael Horowitz, a former Pentagon official now at the University of Pennsylvania, said the OpenAI findings demonstrate how governments may integrate AI tools into routine information operations.
“US-China AI competition is intensifying,” Horowitz said. “It’s not only about who builds the most advanced models. It’s also about how these technologies are used in day-to-day surveillance and influence campaigns.”
Broader implications
The report underscores growing concerns that AI systems can be used not only to generate propaganda but also to coordinate and document complex influence operations.
While OpenAI’s safeguards blocked certain requests, the case illustrates how AI tools — even when not directly producing disallowed content — may still be incorporated into broader campaigns of repression or manipulation.
As AI capabilities expand, experts warn that governments and private companies alike will face mounting pressure to detect and counter the misuse of generative tools in geopolitical conflicts.