A detailed Chinese government secret was accidentally leaked after a Chinese law enforcement official attempted to use ChatGPT to stage a smear campaign against the Japanese prime minister. OpenAI said it banned the account after discovering China’s clandestine strategy of conducting covert influence operations and transnational repression. The Chinese official was using ChatGPT to track their clandestine network and revealed that they were impersonating U.S. immigration officials to intimidate Chinese citizens living in the United States.The Chinese operators apparently posed as U.S. immigration officials to warn Chinese dissidents living in the United States; they also attempted to forge U.S. state court documents to delete a Chinese dissident’s social media account, OpenAI’s report said.
“This is not digitalization, not a prank, but industrialization”
OpenAI principal researcher Ben Nimmo explains that the operation looks industrial. “This isn’t just digital. This isn’t just a spoof. It’s industrial. This is about attacking critics of the Chinese Communist Party. [Chinese Communist Party] Everything, everywhere, happening at the same time. “The Chinese Embassy did not respond to the report. OpenAI doesn’t just follow what officials tell ChatGPT. Investigators matched ChatGPT user descriptions with real-world online activity. For example, the official described how they created a false obituary for a Chinese dissident to spread rumors. And such a thing did happen in 2023.
“Developing a multi-part plan to discredit the incoming Japanese Prime Minister”
The Chinese official asked ChatGPT to develop a multifaceted plan to smear then-incoming Japanese Prime Minister Sanae Takaichi, in part by fanning online anger over U.S. tariffs on Japanese goods. OpenAI said ChatGPT did not respond to prompts. But after the freshman came to power, tags appeared on popular forums for Japanese graphic artists attacking her and complaining about U.S. tariffs.
