OpenAI Bans China-Linked Accounts in Disinformation Probe

OpenAI has banned a series of accounts linked to China, online dating scams, and covert influence operations, including a campaign targeting Japan’s first female prime minister, according to a new transparency report detailing misuse of ChatGPT.

The artificial intelligence company said several accounts used its ChatGPT chatbot alongside social media profiles and other digital tools to conduct cybercrime, impersonate matchmaking agencies and law firms, and pose as US officials.

Among the cases outlined, OpenAI identified a small cluster of accounts believed to originate in China that used its AI models to gather information about American individuals, online forums, and federal buildings. The same accounts also sought guidance on face-swapping software, raising concerns about potential identity manipulation.

According to the report, these accounts sent English-language emails to US state-level officials and political analysts in the business and finance sectors, inviting them to participate in paid consulting opportunities, a tactic often associated with influence operations.

OpenAI said it blocked one ChatGPT account connected to an individual linked to Chinese authorities. That account allegedly coordinated a covert influence campaign aimed at discrediting Japanese Prime Minister Sanae Takaichi, marking a significant case of AI-enabled political interference.

In a separate scheme, a group of accounts used ChatGPT to facilitate a large-scale online dating scam targeting men in Indonesia. OpenAI said the operation likely defrauded hundreds of victims each month by generating promotional materials and advertisements for a fake dating service. Targets were reportedly pressured into completing paid tasks under the guise of platform participation.

Additional accounts were found using OpenAI’s language models to impersonate legitimate law firms and real attorneys, further highlighting the growing sophistication of AI-driven fraud and identity deception.

The company stated that all identified accounts have been suspended and emphasized its ongoing commitment to combating misuse of artificial intelligence technologies.

The findings underscore mounting global concerns about AI-enabled disinformation, online fraud, and cross-border influence campaigns as generative AI tools become more widely accessible.