OpenAI announced in its latest public threat report that it has banned multiple ChatGPT accounts suspected of being linked to Chinese governmental entities, after those users asked the chatbot for plans to build social media “listening” tools and other monitoring systems. 

According to the company, these requests violated its national security policy, since they attempted to exploit generative AI for surveillance purposes. OpenAI also flagged Chinese-language accounts involved in phishing or malware activity, and prohibited users exploring automation via China’s DeepSeek platform.

Beyond China, the platform similarly banned accounts tied to Russian-language criminal groups that were seeking help to craft malware. OpenAI emphasized that its models did not provide novel offensive capabilities to threat actors, but noted that many adversaries use AI to refine existing tactics.

The move underscores growing risks around AI misuse amid geopolitical tensions and raises questions about how to regulate powerful models in the hands of state-linked actors.