Anthropic says AI used in suspected state backed surveillance
Anthropic, an AI company from San Francisco, says its tech was used for suspected state-backed surveillance, and the operations were identified between January and July.
Governments and groups linked to actors in China, Iran, and West Africa reportedly targeted dissidents, ethnic minorities, and activists, like pro-democracy voices in Hong Kong and Tibetan communities, using AI to track people.
Governments replace human monitors with AI
The report highlights a growing trend: governments are replacing teams of human engineers and analysts with AI tools for monitoring.
Iranian actors watched people through social media, while a contractor working for Malian national security authorities used Anthropic's Claude model to develop the software underpinning an intelligence-gathering operation.
Moonshot and DeepSeek allegedly misused Claude
Anthropic also accused Chinese companies Moonshot and DeepSeek of secretly using outputs from Claude to train their own AI models.
Nearly 300,000 customer requests were routed through a network of 5,380 allegedly fraudulent accounts, most of which appeared to originate in Singapore and Japan, raising big questions about privacy.
Anthropic blocked dangerous pathogen research
On the science front, Anthropic blocked efforts to use its AI for research involving dangerous pathogens like avian flu and smallpox.
The company noted, "The individuals implicated in these case studies are working scientists. We do not assert that they intended harm," but still took precautions.