OpenAI's Safety team identified and banned a cluster of ChatGPT accounts linked to a likely China-origin influence operation. These accounts used ChatGPT to generate social media content criticizing US data centers and AI infrastructure. This activity was detailed in OpenAI’s June 2026 report on PRC-linked influence operations targeting AI debates in the US.
The operators prompted ChatGPT in Simplified Chinese to produce English and Chinese-language outputs, posing as Americans from various backgrounds. These outputs were then posted across multiple social media platforms. The accounts accessed OpenAI's platform using VPNs, as direct access from China is not permitted.
Key Points
- OpenAI banned a cluster of ChatGPT accounts likely originating from China.
- The accounts used ChatGPT to generate social media content criticizing US data centers and AI infrastructure.
- This activity was documented in OpenAI's June 2026 report on PRC-linked influence operations.
- Operators prompted ChatGPT in Simplified Chinese for English and Chinese outputs, posing as Americans.
- The campaign primarily targeted US audiences with narratives about data centers increasing electricity demand and costs.
- A secondary audience was overseas Chinese, with attempts to harass Chinese political commentators like Li Ying, Lu Yiheng, Xu Chi, and the X account @SydneyDaddy1.
- ChatGPT was also used for workflow automation, code generation, and text processing tasks.
- The accounts used ChatGPT to generate and refine work reports detailing operational security and platform detection evasion strategies.
Context
According to OpenAI, the operators were likely part of a social media operations team at a private Chinese technology company working for Chinese provincial-level government clients. This activity aligns with a commercial ecosystem supporting Party-state priorities in public opinion guidance. The campaign's objectives and strategies for influencing public opinion and evading platform detection were outlined in a separate report uploaded to ChatGPT by the actors.
Why It Matters
This case demonstrates how AI models can be leveraged in sophisticated influence operations, requiring platform providers to implement robust detection and enforcement mechanisms. It highlights the evolving challenges in identifying and mitigating AI-generated disinformation and coordinated inauthentic behavior.
What To Do
- Note the June 2026 publication date of OpenAI's report for context on the timeline of these observations.
- Observe how AI models like ChatGPT are being used for content generation, workflow automation, and strategic planning in influence operations.
- Watch for further reports from platform providers on AI-assisted influence activities and their mitigation strategies.
- Review the specific examples of content generation and targeting to understand the tactics employed by such campaigns.