The bots doubtless used ChatGPT to generate photos for his or her marketing campaign.
X has found a bot farm community with 200,000 accounts on its platform as a part of its Security staff’s investigation into suspected inauthentic Chinese language accounts. The corporate introduced its findings on its International Authorities Affairs web page. It stated that 200 accounts out of the 200,000 have been posting comedian strips, photos and textual content about AI knowledge facilities in a way that “may manipulate a respectable debate about American AI and power coverage.”
Particularly, their posts talked about how AI knowledge facilities can drive up the costs of family electrical energy and pressure the grids of cities close by. Additionally they posted AI-generated cartoons depicting knowledge heart operators as unscrupulous businessmen raking in income from the amenities whereas common households shoulder the invoice.
X did not checklist the accounts concerned within the troll farm, but it surely did put up some examples on the platform. “We stay dedicated to sustaining an open and genuine platform the place folks debate matters of public curiosity,” it wrote. “We take significantly any makes an attempt to undermine the integrity of the worldwide city sq. and droop accounts that violate our Authenticity coverage.”
Primarily based on the pattern posts X shared, the inauthentic accounts used OpenAI to generate illustrations and speaking factors for his or her marketing campaign. The identical illustrations have been within the report OpenAI printed again in June about how accounts doubtless primarily based in China used ChatGPT to generate anti-data heart content material. OpenAI famous again then that whereas the ChatGPT customers doubtless used pretend accounts to put up on social media, additionally they posted “respectable information tales concerning the energy grid operator’s capability auctions and knowledge heart energy demand.” There are respectable issues about rising energy prices in cities and cities close to hyperscale knowledge facilities. Based on Goldman Sachs, electrical energy costs have already elevated by 6.9 p.c year-over-year as of February 2026, and it believes costs will proceed to rise as AI amenities demand development.
X did not say what the remainder of the accounts within the huge troll farm have been posting about if solely 200 have been concerned in spreading anti-AI knowledge heart sentiments. The report OpenAI printed again in June additionally revealed, nevertheless, that the China-linked ChatGPT customers additionally requested the chatbot to jot down insults in opposition to Chinese language dissidents and political commentators, in addition to criticisms of US tariffs and tech insurance policies.
