WASHINGTON (Reuters) - Chinese propagandists have been trying to use OpenAI's flagship chat to gin up opposition to Donald Trump's tariffs and intervene in American debates over data centers and AI, OpenAI said in a report published on Wednesday.
The company said its efforts, dating to late 2025 and early 2026, appeared to have had little or no effect, but it is another indication of how central generative AI is becoming in digital influence campaigns - even ones aimed against AI companies themselves.
The Chinese Embassy in Washington said that while it was not familiar with OpenAI's research, "we firmly oppose any groundless attacks or smears against China" and that Beijing was working to "ensure AI is a force for good and for all."
OpenAI said it caught one set of Chinese-speaking users using ChatGPT to come up with slogans and cartoons critiquing Trump's trade and tech policy that were later posted to X.
The cartoons featured Trump behaving disruptively on the global stage, for example by swinging a hammer at a wall labeled "Global Future" or sawing apart a ladder he was standing on.
The same group also used the platform to generate Chinese comments for use in the comment sections of Chinese-language articles, as well as content in Italian and Japanese. X's owner, xAI, did not return an email seeking comment.
OpenAI said it had traced another set of users to a Chinese tech company that did government work, although it did not disclose the firm's identity. It said that group tried to intervene in debates over AI and data centers, a hot-button issue in America, where more than a dozen states have or are weighing restrictions on their construction.
Sample cartoons shared by OpenAI painted the industry as greedy profiteers whose electricity consumption was harming ordinary citizens.
Ben Nimmo, a principal investigator for OpenAI, told journalists that the operations seemed geared toward manipulating "a legitimate debate about American AI and wider American tech policies. Under the circumstances it's particularly ironic that they tried to use American AI to do it."
Although the influence operations did not appear to have gone anywhere, the use of AI-assisted propaganda is another illustration of how ubiquitous AI-generated imagery has become, even in propaganda critical of the industry. Last week, the tech publication 404 Media said that digital content farms were using AI-generated imagery to produce memes opposing data centers.