"PRC-linked influence operations are targeting AI debates in the US" is a post published June 10, 2026 on openai.com by OpenAI and tagged Global Affairs. It describes two clusters of ChatGPT accounts, likely originating from China, that OpenAI banned after they used its models "in support of apparent covert influence operations that promoted narratives in an attempt to manipulate a legitimate debate about American AI and wider tech policies."
The post is a summary of a longer June 2026 threat report published separately on OpenAI's content delivery network, which is not held in Raw Sources/. It sits in the same disclosure series as OpenAI's Russia influence-campaign post of August 25, 2026 (Disrupting a new covert influence campaign from Russia).
The two clusters
Data Center Bandwagon. The first cluster "generated social media comments and images claiming that data center buildouts for AI were increasing electricity prices for average families." See AI Data Centers, Data Center Siting / AI Power Politics.
Tech and Tariffs. The second cluster generated comments and images criticising US tariffs "as attempts to dominate technological competition," and specified in its prompts that the content should not include China's leader Xi Jinping in the output and should include only President Trump. OpenAI states this cluster was connected to a network of likely inauthentic social-media accounts that were also likely targeting OpenAI by claiming ChatGPT user data had been compromised, and that "these allegations were entirely false."
OpenAI's stated assessment and its limits
The post is explicit about what it does and does not claim. OpenAI writes that the targeting of the company and of US data-centre buildouts "is significant not because the operation appears to have shifted public opinion, but because it shows PRC-origin influence operators testing narratives against AI infrastructure." It states that the operation "sought to exploit and amplify existing public concerns about energy prices and local impacts of data center development, but we found no evidence of meaningful breakout beyond its own activity."
That limit is load-bearing. The finding is that operators tested narratives, not that the narratives took hold; OpenAI situates the behaviour in a longer pattern in which "foreign influence operations have long sought to latch onto existing local issues and sincerely held beliefs, using them to build credibility, amplify divisions or exacerbate public distrust."
The framing OpenAI gives its own disclosure is mission-linked: that ensuring AGI benefits all of humanity "also requires identifying and disrupting attempts by authoritarian regimes and their proxies to use AI systems to coerce critics, surveil communities or covertly interfere in democratic societies," and that publishing the findings is meant to help industry, governments, civil society and the public identify and disrupt such attempts and resist "totalitarianism with AI characteristics — the use of AI for surveillance, censorship and control over political, social and private life." See AI-enabled influence operations, AI and Authoritarianism.
Later use of the finding
The disclosure was cited in coverage of the US data-centre backlash on September 6, 2026, roughly three months after publication, in reporting that also described technology leaders and the administration attributing the populist movement against AI to foreign influence (Source: wsj.com). OpenAI's own no-breakout finding does not support the broader attribution the later argument makes; the post claims operators tested narratives against an existing debate, not that the debate originated with them.
Provenance
Captured in full from openai.com, OpenAI's own domain and canonical host for its threat-intelligence disclosures, on September 7, 2026. Tagged Global Affairs, bylined "OpenAI." The campaign names and the prompt instruction to exclude Xi Jinping appear verbatim.
Two related documents were not captured. The underlying full threat report is a separate PDF at cdn.openai.com and covers the campaign detail this summary omits. A companion Global Affairs post on X identifying roughly 200 China-linked anti-data-centre accounts sits behind an authentication wall.
The date requires care. The post's own cover date is June 10, 2026; it entered the record through September 2026 coverage that cited it without a date, in a context reading as a contemporaneous disclosure. Any claim drawn from it dates the disclosure to June and the coverage to September.
Relationships
- supports: AI-enabled influence operations
- related: Disrupting a new covert influence campaign from Russia — companion disclosure in the same series
- related: AI Data Centers
- related: Data Center Siting / AI Power Politics
- related: OpenAI