Anthropic Disrupts AI Espionage Originating from China, Iran, and West Africa
The technology company announced it has neutralized state-backed artificial intelligence operations targeting dissidents and minorities.
Anthropic announced that between January and July, it stopped multiple state-backed surveillance operations based in China, Iran, and West Africa that misused artificial intelligence models to target dissidents and ethnic minorities.
Targets of Surveillance Operations
The published report stated that the detected espionage campaigns aimed at diasporas and opposition communities historically targeted by regimes. In this context, pro-democracy figures in Hong Kong, Tibetan and Falun Gong communities in Asia, as well as Iranian minorities and anti-regime figures abroad, were targeted.
AI-Powered Tactics
It was reported that Iranian actors developed a method to identify people via social media accounts, while a contractor working for Mali National Security officials used the Claude model to design intelligence-gathering software.
Model Theft and Covert Redirection
The San Francisco-based AI laboratory reported that Chinese developers deceptively used Claude to generate responses to user queries and covertly transferred this data to develop their own models.
Biological Research and Weapon Attempts
Anthropic also stated that it blocked users' attempts to design weapons, conduct suspicious biological research, and create romance scams.