Misuse and incidents

China-linked actors use ChatGPT to develop protest-surveillance tooling

OpenAI disrupted likely China-origin accounts that used ChatGPT to analyze political material, draft sales pitches and debug code for a social-listening tool claimed to feed real-time protest reports to Chinese security services.

CURRENT ASSESSMENT · REVISION 1
TOWARD DOOM42confidence 85/100

Why it moved the index

The case directly connects general-purpose model assistance to attempted authoritarian communications surveillance and protest monitoring. Confidence is lower because the operators' claim of operational reporting could not be fully verified and the core listening tool used a non-OpenAI model.

0 comments · 0 votesOpen discussion

Public discussion is readable by everyone. Sign in to comment, reply, or vote.

No comments yet. Start the discussion.

AUDIT TRAIL

Assessment history

  1. R1
    Toward 42 · confidence 85

    New February 2025 surveillance-misuse case materially distinct from the same report's influence operation and absent from durable records.

    12 Aug 2026