OpenAI has banned accounts linked to the PRC government that were using ChatGPT to support surveillance, profiling, and research on critics. The company emphasized its commitment to democratic AI principles, aiming to prevent authoritarian regimes from using AI to control citizens.

These findings, based on a selection of disrupted activities, offer insights into how such regimes might misuse AI. Some banned users sought to develop tools for large-scale online monitoring, analyzing data from social media platforms. One user requested assistance in creating promotional materials and project plans for a social media listening tool, purportedly for a government client, designed to scan platforms for specific content.

Other banned users appeared to be engaged in targeted profiling and research. Examples include requests to identify funding sources for a critical X account and to find organizers of a petition in Mongolia. While OpenAI's models only provided publicly available information, this activity highlights the potential for AI misuse in surveillance and control.