OpenAI recently uncovered and dismantled two clusters of ChatGPT accounts linked to China that were attempting to manipulate American public opinion. These covert operations, detailed in the company’s June 2026 threat report, sought to inject false narratives into sensitive domestic debates. The campaigns specifically targeted the rapid expansion of artificial intelligence infrastructure in the United States and ongoing discussions regarding international trade tariffs.
The first operation, which OpenAI internally dubbed the “Data Center Bandwagon” campaign, focused on exploiting public frustration over energy costs. Operators used ChatGPT to generate social media comments and images claiming that the construction of large-scale AI data centers is directly responsible for rising electricity bills for average families. By posing as concerned American citizens on platforms like X and Facebook, these accounts aimed to amplify anxiety surrounding energy consumption and the local impact of new tech developments.
A second cluster, labeled the “Tech and Tariffs” campaign, took aim at US economic policy. These accounts generated content criticizing US tariffs, portraying them as aggressive attempts to dominate global technological competition. Notably, the operators explicitly instructed ChatGPT to include President Donald Trump in their generated imagery while carefully excluding China’s leader, Xi Jinping. This cluster also linked to a network of inauthentic accounts that spread entirely false claims suggesting that ChatGPT user data had been compromised, an allegation that OpenAI has categorically denied.
OpenAI’s investigation traced these activities to individuals likely based in China, specifically at a private technology company that holds multiple contracts with provincial-level government clients. The operators utilized VPNs to bypass regional access restrictions and prompted the AI in simplified Chinese to produce output in both English and Chinese. Despite these efforts, the report indicates that the campaigns were relatively small in scale. OpenAI’s internal “Breakout Scale” rated these operations at Category One, the lowest possible impact level, noting that they failed to generate any meaningful authentic engagement beyond their own networks.
While the immediate impact of these campaigns appears minimal, the situation highlights a growing trend of foreign adversaries testing AI tools to exacerbate political divisions. Ben Nimmo, the principal investigator on OpenAI’s Intelligence and Investigations team, described the activity as a classic example of a foreign operation “jumping onto the bandwagon” of pre-existing domestic debates. By using American-made AI to stir controversy within the United States, these actors demonstrated a strategic intent to weaponize existing societal tensions.
The disclosure has triggered significant attention from Washington, where policymakers are increasingly concerned about foreign influence in the tech sector. Representative John Moolenaar, chairman of the House Select Committee on the Chinese Communist Party, emphasized that while Americans have legitimate questions about the energy demands of data centers, the Chinese government is actively exploiting this open discourse to divide the public. As AI infrastructure becomes a cornerstone of national economic competition, the incident serves as a clear warning about the evolving risks posed by AI-powered disinformation in the digital age.
