OpenAI says China-based actors used ChatGPT in influence operation targeting US AI data centres

OpenAI has said it banned a cluster of accounts it believes were likely based in China after finding they used ChatGPT in a covert influence operation aimed at stoking opposition to AI data centres in the United States. The company said the activity was designed to exploit existing public concerns about electricity prices and to manipulate debate about American AI. It also said a separate cluster of accounts produced material about US tariffs and competition with China.

Sponsored

In a research report released on Wednesday, OpenAI said the accounts generated social media comments and images blaming data centres for rising electricity bills in communities across the US. One example cited by the company was a comic strip showing a cigar-chomping businessman holding bags marked with dollar signs as a family reacted in shock to its electricity bill. OpenAI said the material was created to appear as part of a legitimate public discussion while concealing who was behind it.

The company said it found no evidence the campaign had a meaningful influence. Even so, it described the activity as an attempt to insert itself into an ongoing American debate about the future of the country's AI capabilities. OpenAI said foreign influence operations often seek to latch on to local issues and sincerely held beliefs in order to build credibility, amplify divisions or deepen distrust.

The disclosure matters because AI data centres have become a politically sensitive issue in the US, where concerns about energy use and electricity costs are already part of wider debates over the expansion of artificial intelligence. OpenAI's findings suggest those arguments can be targeted by actors seeking to shape public opinion through synthetic content. The report also points to the growing use of generative AI tools in information operations, including the rapid production of text and images tailored to specific political themes.

Sponsored

OpenAI said the second cluster of accounts generated content casting US tariffs as an effort to dominate technological competition with China. It added that the material was instructed not to mention Chinese leader Xi Jinping. The company said the two clusters appeared to be part of separate efforts, but both were aimed at influencing how users viewed US policy and technology competition.

China's embassy in Washington said it was not familiar with the report, but opposed what it called groundless attacks or smears against China. OpenAI said the operators tried to covertly enter an American debate while hiding who they were and what motivated them. What remains unclear is how widely the content spread, whether any real-world audiences were persuaded, and whether similar campaigns are still active.

The company's report adds to scrutiny of how generative AI systems can be used to scale influence efforts, and what safeguards may be needed as those tools become more widely available.

#OpenAI #ChatGPT #China #influenceoperation #AIdatacentres

Sponsored