OpenAI Uncovers China-Linked Influence on US AI Data Debate

OpenAI announced that it had identified two clusters of accounts it believes originated from China that used its ChatGPT model to generate large volumes of social‑media content aimed at shaping U.S. discussions about artificial‑intelligence data centres and U.S. tariff policy. The “Data Center Bandwagon” campaign produced posts, comments and political cartoons arguing that AI data centres were driving up electricity costs for ordinary American families. The “Tech and Tariffs” campaign created slogans and cartoons critiquing former President Donald Trump’s trade measures, explicitly directing the model not to mention Chinese President Xi Jinping.

OpenAI said the accounts were operated through likely inauthentic profiles on platforms such as X and were subsequently banned. The company found no evidence that the campaigns achieved meaningful engagement or altered public debate. U.S. Representative John Moolenaar, chair of the House select committee on the Chinese Communist Party, warned that the operation illustrated how China seeks to exploit America’s open political system. The Chinese embassy in Washington denied familiarity with the report and rejected what it described as unfounded smears.

OpenAI stressed that while the influence attempts were small‑scale, they highlight a growing risk that generative AI tools can be weaponised for foreign influence operations, even against the companies that provide the technology.