OpenAI bans China-linked accounts for using ChatGPT in US influence campaigns

OpenAI has shut down a cluster of China-linked accounts that were using ChatGPT to craft social media content designed to shape American public opinion. The operation, which the company has internally labeled “Uncle Spam,” focused on hot-button issues like US tariffs and AI data center infrastructure.

OpenAI rated the effort as Category 2 on its influence operations scale, meaning engagement across platforms like X and Bluesky was negligible.

What OpenAI found and shut down

The banned accounts were part of a broader pattern that OpenAI has been tracking for months. Over a three-month period leading up to mid-2025, the company says it disrupted at least 10 malicious operations. Four of those were specifically tied to actors linked to the Chinese government.

Accounts used ChatGPT to generate social media posts, develop fake personas, and scrape data. Some probed the platform’s guardrails around politically sensitive topics, essentially testing where the lines were drawn before trying to cross them.

OpenAI had flagged similar activity earlier this year. The company disclosed comparable enforcement actions in both February and October 2025, targeting Chinese-linked operations involved in influence campaigns and, in some cases, malware distribution.

The bigger picture on AI and information warfare

The low engagement numbers suggest that creating convincing AI-generated content is only half the battle. Distribution, amplification, and building authentic-looking networks still require significant human effort or pre-existing infrastructure.

Russian-linked entities have also been flagged in OpenAI’s threat reports, suggesting that multiple state actors are running parallel experiments with generative AI as a propaganda tool.

What this means for investors

No tokens, blockchain protocols, or digital asset platforms were implicated in the “Uncle Spam” operation or any of the other disrupted campaigns.

The fact that AI data centers were one of the targeted topics in these influence campaigns reflects how politically charged the buildout of AI computing infrastructure has become. Energy consumption debates have historically spilled over into crypto, particularly around proof-of-work mining.

OpenAI’s willingness to publicly name and disrupt state-linked operations signals that AI companies are positioning themselves as active participants in national security conversations.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.