OpenAI dismantles ten covert influence operations with ChatGPT
OpenAI recently revealed that it has shut down 10 covert operations. These would involve propaganda using artificial intelligence. Four of these campaigns OpenAI attributed to China.
Information Security Netherlands June 11, 2025
"What we're seeing from China is a growing number of covert operations using an increasingly broad range of tactics," says Ben Nimmo, principal investigator of OpenAI's intelligence team.
Fake involvement
In these operations, they used ChatGPT to create posts on social media and fake engagement. They also created internal performance reviews that documented their activities on platforms such as TikTok, Facebook, Reddit and X.
Journalists and commentators
In one of these Chinese operations, actors posed as journalists and geopolitical commentators. Using ChatGPT, they wrote account biographies and translated communications between Chinese and English. They also analyzed correspondence addressed to a U.S. senator about a federal nomination. The actors used models from OpenAI to create marketing materials. These included ads for fake social media campaigns and recruiting intelligence sources.
Click
here for the post on Perplexity.
Click
here for NPR's message.
