Generative AI in transnational repression cases and the response from AI companies
Anthropic's September threat intelligence report describes state-linked actors using Claude for surveillance, some of it aimed at dissidents living abroad. In an October commentary for Just Security, Citizen Lab researchers analysed the report and urged AI companies to act on disclosure and notification, detection and refusal, and evidence preservation. Neither document offers settings advice for ordinary users.
Based on 8 sources, including Anthropic, Just Security, Citizen Lab