OpenAI Links Chinese Influence Operation to Anti-Data Center Campaign Inside the U.S.

OpenAI disclosed Wednesday that influence operators likely based in China used ChatGPT accounts to shape American public opinion on artificial intelligence and technology policy, including a coordinated effort to stoke opposition to data centers on U.S. soil.

In a report published the same day, the company said it identified two clusters of ChatGPT accounts "likely originating" from China. One of those clusters generated social media comments and images designed to amplify concerns about American AI infrastructure — including narratives targeting data center construction.

The disclosure arrives as data centers have become a flashpoint in domestic policy debates, underscoring how foreign actors have moved to exploit existing domestic tensions over the energy, water, and land demands that large-scale AI facilities place on local communities.

The findings also land against a backdrop of heightened regulatory scrutiny. Texas Gov. Greg Abbott on Wednesday directed the Texas Public Utility Commission and the Electric Reliability Council of Texas — the state's grid operator — to "take immediate steps" to safeguard the grid from data center strain. Abbott's directives include proposals to require water-efficient technology at data centers and to repeal existing tax incentives that have drawn large facilities to the state.

Whether the Chinese influence operation had any measurable effect on state-level policy deliberations remains unclear. OpenAI did not attribute specific policy outcomes to the accounts it identified.

The timing of the two developments — OpenAI's report and Abbott's announcement, both on Wednesday — reflects the degree to which data center policy has become a contested terrain, drawing in state regulators, federal policymakers, and, according to OpenAI, foreign influence networks simultaneously.

Adding another layer to the week's policy turbulence, President Trump suggested Thursday that he could use an executive order to extend Section 702 of the Foreign Intelligence Surveillance Act after Congress failed to pass a temporary extension of the surveillance program. Section 702 authorizes the collection of foreign intelligence communications and has been a central tool in monitoring foreign influence activities.

"Congress wants me to do it," Trump said in the Oval Office, when asked by a reporter whether an executive order was under consideration. "Let's see what happens," he added.

The potential lapse or executive extension of Section 702 carries direct relevance for counterintelligence operations of the kind that would track influence campaigns like the one OpenAI described. Without the authority, intelligence agencies face constraints on how they monitor foreign actors operating through commercial platforms.

OpenAI did not specify which social media platforms were targeted by the Chinese-linked accounts, nor did it disclose the total number of accounts involved in either cluster. The company also did not say whether it had shared its findings with federal law enforcement or intelligence agencies.

The confluence of OpenAI's disclosure, Abbott's regulatory push, and the Section 702 debate signals that AI infrastructure — its physical footprint, its policy environment, and its vulnerability to manipulation — has become a central front in a broader contest between the United States and China over technological influence.