Chinese law enforcement used ChatGPT for cross-border intimidation and disinformation
OpenAI reported that a Chinese law enforcement official used ChatGPT to plan and document covert influence operations, including intimidating overseas dissidents, spreading disinformation about Japanese Prime Minister Sanae Takaichi, and creating fake death announcements of activists.
What happened
A Chinese law enforcement official used ChatGPT to plan and document covert influence operations, according to OpenAI's investigation. The official posed as U.S. immigration officials and used forged U.S. court documents to threaten Chinese dissidents living in the United States, warning them that their public statements were illegal.
The official created fake funeral announcements and tombstone photos to spread false reports of an activist's death. The official also requested ChatGPT to develop plans to discredit Japanese Prime Minister Sanae Takaichi by inciting Japanese public anger over U.S. tariffs, though ChatGPT refused that request.
OpenAI found that ChatGPT was used as a tool to document secret online operations, with content generated by other tools and distributed through multiple social media accounts and websites. OpenAI terminated the account after discovering the activity.
Reported impact
Japanese internet forums displayed attack tags and complaints about U.S. tariffs following Prime Minister Takaichi's appointment in October 2025, correlating with the documented campaign plans.
The operations represented industrialized coordinated activity targeting critics of the Chinese government across multiple platforms and locations.
Index analysis
OpenAI's investigation showed that the platform was used to facilitate influence operations at scale rather than isolated misuse. The account holder used ChatGPT alongside social media accounts and other tools to execute coordinated disinformation and intimidation campaigns across borders.
Governance controls
OpenAI banned the ChatGPT account linked to the Chinese law enforcement official. The company detected and shut down the account after investigating the activity.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.