Spamouflage Chinese influence operation used LLMs to draft posts
What they were trying to do
- 01
Generate short social-media comments in multiple languages for cross-platform distribution
- 02
Debug and improve scripts automating coordinated posting across social platforms
How they did it
Spamouflage used OpenAI models to generate short multilingual social-media comments for cross-platform distribution and to debug automation scripts used in the posting workflow. The operation bolted AI into an existing scripted distribution infrastructure rather than replacing it.
Example queries
Verbatim excerpts from the disclosure — only what the vendor actually published.
Not detailed in the disclosure
What they achieved
Concrete outputs obtained before platform disruption.
- before disruption
Short social-media comments in multiple languages generated for distribution
- before disruption
Debugging assistance for cross-platform posting automation scripts
OpenAI banned the associated accounts. OpenAI assessed the operation had not succeeded in meaningfully building audience engagement before disruption.
Full summary
PRC-linked Spamouflage (DragonBridge) operation used OpenAI models to generate short social-media comments in multiple languages and to debug scripts that automated cross-platform posting. OpenAI banned the accounts.
AI platforms involved
Related incidents (same country or vendor label)
- CN · China
GTG-1002: first publicly disclosed AI-orchestrated cyber espionage campaign (PRC state-sponsored)
2025-11-13Anthropic - CN · China
APT41 and 20+ PRC-affiliated clusters used Gemini for post-compromise scripting and EDR reverse engineering
2025-01-29Google (GTIG) - CN · China
SweetSpecter spear-phished OpenAI employees while using ChatGPT for recon
2024-10-09OpenAI - CN · China
Charcoal Typhoon used LLMs for tooling and social-engineering research
2024-02-14Microsoft + OpenAI - CN · China
Salmon Typhoon used LLMs for translation and intel-agency research
2024-02-14Microsoft + OpenAI
Source note: Every entry in this catalog is derived from a primary public disclosure by the named platform vendor. This entry cites OpenAI as the primary source. No private, speculative, or non-public claims are presented here.