OpenAI Identifies China-linked Influence Campaigns Targeting Tariffs, Data Centers

According to OpenAI's report published June 10, 2026 and reported by Reuters, Axios, Cyberscoop and other outlets, OpenAI's threat intelligence team identified two distinct clusters of activity that likely originated in China and used ChatGPT to generate social media content aimed at U.S. debates on tariffs and AI data centers. OpenAI named the two operations "Data Center Bandwagon" and "Tech and Tariffs," per Axios. The campaigns produced slogans, political cartoons and comment text in simplified Chinese and English, and requested outputs in other languages including Italian and Japanese, per Cyberscoop and Reuters. OpenAI traced "Data Center Bandwagon" to an unnamed Chinese technology company that does government work, Reuters reports; that group asked ChatGPT to create comic strips about power grid capacity and electricity prices, which were later posted to X via likely inauthentic accounts alongside links to legitimate news stories. The "Tech and Tariffs" group generated cartoons criticizing Trump's tariff and tech-dominance policies; per Axios, the prompts directed ChatGPT to depict Trump but not Xi Jinping. The company says the activity dates to late 2025 and early 2026. The reporting notes the operations appear to have used VPNs and fake American personas, and that the campaigns showed limited reach outside their own amplification networks. Ben Nimmo, principal investigator at OpenAI, is quoted by Reuters saying, "This looks like a classic example of a foreign influence operation jumping onto the bandwagon of a genuine and pre-existing domestic debate and trying to manipulate it by using fake accounts posing as Americans." Nimmo told reporters separately, per Axios: "The debate existed already. This was an influence operation from China trying to interfere in it." The Chinese Embassy in Washington said it was not familiar with the research and, according to Reuters, "we firmly oppose any groundless attacks or smears against China."

Actors repurposing generative models for influence operations is consistent with recent trends in low-cost information operations. Public reporting describes the campaigns using repeated prompting in simplified Chinese, multi-language outputs, AI-generated imagery and VPN routing to mask origins. These tactics lower the marginal cost of producing large volumes of tailored, local-language content and make early detection harder because outputs can be novel and varied. Industry detection signals reported in coverage include cross-account coordination patterns, repeated prompting, and amplification inside closed networks rather than broad organic spread.

Observed patterns in similar transitions show platforms, researchers and national security teams increasingly face cross-border influence tests that latch onto genuine domestic debates rather than invent issues from scratch. Public reporting frames this episode as an example where generative AI acts as an accelerant for content creation, not as the originator of the underlying grievance. For practitioners building moderation, threat-detection or forensic tooling, flagged attributes from the coverage, multilingual outputs, VPN use, repeated prompting and imagery generation, are practical signals to incorporate into models and pipelines.

For practitioners and platform security teams, industry observers will watch for:

• •increases in cross-language meme and cartoon campaigns that mirror local arguments
• •broader use of commercial LLMs in coordinated influence activity
• •how platforms disclose and share indicators of coordinated inauthentic behavior. Policymakers and state actors may also escalate scrutiny of data center siting debates cited in the reporting, especially where private contracts or regional infrastructure projects are named in public filings. Reporting does not provide evidence that the campaigns materially shifted public opinion, and OpenAI's materials indicate limited downstream amplification in this instance