ChatGPT Blocked Chinese Cop's Plot to Smear Japan's PM

OpenAI dropped a notable disclosure: ChatGPT caught and blocked an attempt by someone tied to Chinese law enforcement to weaponize the AI chatbot for political disinformation.

The user reportedly tried to get ChatGPT to help plan an online campaign designed to discredit Japanese Prime Minister Sanae Takaichi. The system refused.

It's a concrete example of AI guardrails actually working in the wild — and against a state-affiliated actor, no less. OpenAI has been ramping up its threat detection capabilities as governments and bad actors increasingly probe AI tools for influence operations.

The incident highlights the growing intersection of AI, geopolitics, and information warfare. Foreign influence campaigns aren't new, but using commercial AI tools to scale them is an evolving threat that companies like OpenAI are being forced to police in real time.