OpenAI, the US company that operates ChatGPT, released a report saying that a user associated with Chinese law enforcement tried to use the conversational artificial intelligence to plan a covert influence operation aimed at defaming Japanese Prime Minister Takaichi Sanae.

According to the report, the attempt occurred around the time of Takaichi’s appointment as prime minister in October 2025.

The user tried to use fake social media and email accounts to spread comments that undermine Takaichi’s credibility on issues such as rising prices and policies on foreign nationals, but ChatGPT refused to cooperate.

OpenAI said it may restrict usage when its terms of service are violated.

According to the company’s analysis, the influence operation involving spreading defamatory images and comments about Takaichi was carried out using other AI services, but no significant impact has been observed.

OpenAI said the account in question has been suspended.