Join The Club
ChatGPT developer OpenAI has published a new security report alleging that, in essence, its own tools are being deployed against it. The company's June 2026 threat report is titled "PRC-linked influence operations are targeting AI debates in the US," and claims that China-based actors are using ChatGPT to whip up anti-data-centre, anti-tariff, and anti-US sentiment online.
OpenAI says it has "banned a cluster of ChatGPT accounts that likely originated in China and used ChatGPT to generate social media content for a covert influence operation." What were they up to? Well, generating a lot of bad AI political cartoons, for one thing.
These users would—writing in simplified Chinese, the standard form of the written language in mainland China—ask ChatGPT to generate political cartoons that took aim at, for instance, spiking electricity costs caused by AI data centres and Donald Trump's vindictive behaviour toward notional American allies.
The tech was also used to generate antisemitic memes about "Jewish capital" dictating American policy, and to besmirch Chinese dissidents. The relevant prompts "repeatedly used terminology consistent with individuals associated with China’s public security system," says OpenAI.
Cartoons, short phrases and rumours of an OpenAI data leak that never happened were then shared on social media like X and Facebook by networks of fake accounts.
"It is ironic," says OpenAI, that the scheme "used American AI, rather than Chinese models, to generate their content about American AI. We are not in a position to determine what drove this choice".
Now, before we get too ahead of ourselves, OpenAI is not actually claiming that the reason people don't like data centres is because they've fallen prey to Chinese influence operations. Indeed, the report states and restates that the efforts it claims to have uncovered achieved, well, basically nothing.
"Using the Breakout Scale, we assess this activity as Category One: activity spanning one platform, with no evidence of breakout," writes OpenAI in its impact summary. OpenAI also admits that, well, a great deal of the material these efforts drew on was entirely legitimate reporting about the impact of data centres and the blowback from US tariffs.
Which does raise the question: does any of this really matter? OpenAI certainly thinks it does. These attacks "attempted to connect US technology policies and industries to everyday economic anxieties and geopolitical instability," says the company, and show the potential for "influence operations originating from China" to be "inserted into legitimate public debates while nudging audiences toward distrust of US institutions, technology companies and democratic policy choices to help Beijing gain a strategic advantage in AI development".
One might question whether American citizens protesting against data centre construction in their neighbourhoods feel that they were adequately democratically consulted on the whole thing, but that's OpenAI's line and it's sticking to it.
It's also the case that, while OpenAI's report might say the impact of these campaigns was negligible, that's not what everyone says. Pro-data-centre politicians in the US are already seizing on a narrative that anti-AI sentiment is being driven by shadowy foreign interference, and reports like this one—regardless of their actual conclusions—will fuel that story.
In a comment to Reuters, the Chinese Embassy in Washington said that it had not read OpenAI's report, but that "we firmly oppose any groundless attacks or smears against China." China, said the diplomats, wants to "ensure AI is a force for good and for all".
