A Chinese judicial official is suspected of attempting to use the artificial intelligence (AI) chatbot ChatGPT to run an influence operation targeting Japan's prime minister. However, analysts said the actual impact appeared limited.
According to OpenAI's report on blocking malicious AI use released on the 26th, a person believed to be affiliated with a Chinese law enforcement agency planned and documented a coordinated smear operation against Prime Minister Sanae Takaichi using ChatGPT in mid-October last year.
The individual labeled it a "cyber special operation" and sought to post negative comments about Prime Minister Takaichi, send critical emails posing as foreigners, and engineer public sentiment to stoke resentment against U.S. tariff measures on Japan.
OpenAI said ChatGPT refused requests for advice on the operation. But it added there was a possibility the operation proceeded without ChatGPT's help, citing that the same user later asked for edits to a document with similar content.
In fact, on social media platforms such as X (formerly Twitter), posts and YouTube videos critical of Prime Minister Takaichi using the hashtag "right-wing symbionts," which appeared in the document, were reported to have been uploaded.
However, OpenAI assessed the influence as limited, noting that related YouTube videos drew only single-digit views and most social media posts failed to gain much traction.
The report also said Chinese judicial authorities have employed more than 100 tactics, including producing and distributing fake obituaries and tombstone photos of dissidents and conducting activities targeting human rights groups. It added that in the process, not only ChatGPT but also domestic AI models such as DeepSeek appeared to have been used.