Are AI tools becoming the new battleground for geopolitical influence campaigns? OpenAI report

OpenAI has reported that it detected two distinct clusters of accounts using ChatGPT that it believes were linked to China and involved in coordinated influence operations. According to the company, these accounts were used to generate social media content and images aimed at shaping public debate around US artificial intelligence infrastructure and technology policy.

The company stated that while the activity did not show evidence of significantly influencing broader public opinion, it appeared to be an attempt to test narrative strategies around sensitive technology and policy issues.

First cluster: “Data Center Bandwagon”

OpenAI identified the first cluster as part of what it termed the “Data Center Bandwagon” campaign.

This group allegedly generated posts and images suggesting that the expansion of AI data centres in the United States was driving up electricity costs for ordinary households. The content was designed to align with existing public concerns about rising energy consumption and infrastructure demands linked to artificial intelligence growth.

According to OpenAI, the activity appeared to amplify debates around the environmental and economic impact of large-scale AI infrastructure development.

Second cluster: “Tech and Tariffs”

The second cluster, labelled the “Tech and Tariffs” campaign, was said to have focused on US trade and technology competition narratives.

OpenAI stated that this group produced content criticising US tariffs and framing them as efforts to dominate global technological competition. The prompts reportedly included instructions to exclude references to China’s leadership while focusing on US President Donald Trump.

The company also said this network was linked to inauthentic social media accounts that spread false claims suggesting ChatGPT user data had been compromised. OpenAI described these allegations as entirely false.

Nature of the operation

OpenAI said the activity appeared to be part of an attempt to insert itself into ongoing political and policy debates in the United States, particularly those related to artificial intelligence infrastructure, energy usage, and global technology competition.

The company noted that the operations did not appear to achieve significant reach or influence beyond their own coordinated activity. However, it emphasised that such campaigns demonstrate efforts by foreign actors to test how AI tools and narratives can be used in shaping public discourse.

Broader concerns raised by OpenAI

OpenAI highlighted that the targeting of AI infrastructure debates is significant because these systems form a core part of technological and economic development in the United States.

It warned that such influence operations may aim to exploit existing public concerns around energy prices, infrastructure development, and economic competition. The company said it is working to help governments, industry and civil society identify and counter attempts to manipulate legitimate public discussions.

OpenAI also referenced concerns about what it described as attempts to advance “authoritarian-style” uses of AI, including surveillance, censorship and social control, though it did not provide independent verification of broader geopolitical claims.

OpenAI has not indicated that the identified activity had a measurable impact on public opinion. It stated that the accounts appeared to be part of experimentation with narrative strategies rather than large-scale influence success.

The company continues to monitor and analyse such activity as part of its broader efforts to detect misuse of AI systems for coordinated information operations.