An alleged influence op against Japan PM Sanae Takaichi has been exposed after OpenAI revealed it banned a ChatGPT account linked to Chinese law enforcement. The artificial intelligence company said the user attempted to use its chatbot to plan a coordinated online campaign targeting the Japanese leader.
According to OpenAI’s latest threat report, the account activity suggested a wider, well-funded strategy aimed at both domestic and foreign critics. The attempted influence op against Japan PM Takaichi included requests for negative content, fake accusations and social media messaging designed to stir public anger.
OpenAI stated that its system refused to generate the requested materials. However, further activity indicated that the operation continued independently of the platform.
Influence Op Against Japan PM Uncovered in Threat Report
OpenAI said it detected suspicious activity in mid-October after Prime Minister Sanae Takaichi criticised human rights conditions in China’s Inner Mongolia region. The user linked to Chinese law enforcement allegedly asked ChatGPT to draft hostile narratives about the Japanese leader.
These requests included negative comments about Takaichi, fabricated complaints about immigration and the cost of living, and claims suggesting far-right leanings. The user also sought content designed to inflame anti-US sentiment over tariffs and to portray Inner Mongolia in a positive light.
When the chatbot declined to assist, the individual later returned to ask ChatGPT to edit a status report on the campaign. This suggested that the influence op against Japan PM Takaichi had progressed using other methods.
Hashtag Campaign and Limited Engagement
The reported operation included coordinated hashtag use across multiple platforms. One hashtag referenced the far right and appeared alongside anti-Takaichi and anti-US memes.
OpenAI said it identified more than 50,000 posts across over 200 Western platforms. However, fewer than 150 of those posts received more than 300 shares or comments.
Despite the scale of activity, OpenAI concluded that the influence op against Japan PM Takaichi did not achieve significant engagement or measurable impact.
Japanese media later reported on a broader disinformation campaign targeting Takaichi and her government. Research cited by local analysts indicated that thousands of social media accounts were involved, spreading fabricated claims in Japanese and English, some using AI-generated images.
Wider Pattern of Cyber Operations
OpenAI’s findings suggested that the influence op against Japan PM Takaichi was part of a larger strategy. The user documented what were described as “cyber special operations” aimed at suppressing dissent and intimidating critics.
According to the report, these operations relied on extensive resources. They allegedly involved hundreds of staff, thousands of fake social media accounts and coordinated tactics such as trolling, impersonation and mass posting.
The chatbot logs also referenced activity on Chinese platforms such as Weibo and WeChat, alongside hundreds of foreign networks. OpenAI said the user described millions of posts on Chinese social media and tens of thousands internationally.
AI Tools Used for Monitoring and Content Creation
The threat report highlighted how AI was allegedly used within the broader operation. The user reportedly described AI tools being deployed for monitoring, profiling, translation and content creation.
OpenAI said this reflects a systematic use of artificial intelligence to amplify influence campaigns and document internal progress.
In addition to the alleged influence op against Japan PM Takaichi, OpenAI said it banned other accounts linked to separate malicious activities. These included AI use in fraudulent schemes targeting young men in Indonesia.
Global Concerns Over AI-Driven Disinformation
The exposure of the influence op against Japan PM Takaichi underscores growing concerns about AI-powered disinformation. As generative tools become more accessible, governments and technology companies face increasing pressure to detect and prevent misuse.
OpenAI emphasised that its systems refused to generate harmful content and that it continues monitoring suspicious behaviour. The company’s periodic threat reports aim to shed light on emerging tactics used in coordinated online campaigns.
While the alleged operation appears to have had limited impact, experts warn that sustained efforts using AI could evolve in sophistication.
The incident adds to broader geopolitical tensions over digital interference and the weaponisation of online platforms. As election cycles and political disputes intensify worldwide, the battle against AI-enabled disinformation is likely to remain a critical challenge.
