Anthropic says governments used Claude for surveillance
Anthropic says state-linked actors used Claude to automate parts of surveillance work, pushing AI misuse from theory into operational reality.
Commercial AI models are increasingly being used for tasks far beyond productivity, including surveillance workflows that once required much more human labour.
What happened
Anthropic disclosed cases where actors linked to governments used Claude to automate parts of surveillance operations, including gathering information, analysing social-media activity and producing intelligence reports. Anthropic said it banned the accounts it identified.
Why it matters
AI can lower the cost and effort required to monitor large numbers of people, making surveillance more scalable even for smaller teams. That changes the economics of state and non-state intelligence work.
The bigger picture
AI misuse is becoming a cybersecurity and governance issue, not just a model-safety problem. As general-purpose models become more capable, providers will face increasing pressure to detect and restrict harmful operational use.