China, Iran among countries that have used AI to aid spying, Anthropic says

SAN FRANCISCO — Anthropic disclosed on Thursday that it had terminated several instances of state-backed surveillance operations that exploited its artificial intelligence (AI) models. The report, detailing the misuse of its systems, revealed that these activities were carried out by nations and state-aligned actors, primarily between January and July. The perpetrators were located in China, Iran, and west Africa, according to Anthropic's investigation.

The surveillance campaigns primarily targeted diaspora and dissident communities that these governments have historically monitored. These included pro-democracy activists in Hong Kong, Tibetan and Falun Gong communities across Asia, as well as Iranian minority groups and opponents of the Iranian regime residing abroad. The report explained that these regimes had been utilizing AI to spy on these communities for some time.

One instance involved Iranian actors devising a method to identify individuals through their social media accounts. In another case, a contractor employed by Malian national security authorities leveraged Anthropic's AI model, Claude, to create the software facilitating intelligence gathering. The report noted that AI is increasingly replacing the need for a dedicated engineering workforce in such surveillance operations.