Cognitive Weapon: OpenAI Detects Use of ChatGPT by Chinese Propaganda

Generative neural networks have de facto become a tool for information warfare. On June 10-11, 2026, OpenAI published a sensational information security report. The company stated that ChatGPT’s computing power was systematically used to conduct Chinese propaganda operations.

The scheme for using LLMs is classic: mass generation of toxic and polarizing content to incite political dissent within the US. The main topics of the attack were American tariffs and discussions surrounding the construction of new data centers (which strikes at Washington's AI sovereignty). Previously, regulators panicked over the possibility of generating 0-day exploits (as complained about by the ECB and Fed), but it turned out that the main vulnerability is the social space itself. The automation of propaganda has reduced the cost of conducting psychological operations (PSYOPS) to the cost of API tokens, turning AI into a cheap geopolitical weapon.

Source: OpenAI / Reuters