Anthropic Is Building a Predictive Surveillance System to Monitor Activists
20 days ago
- Anthropic is building a comprehensive monitoring system to track activists who oppose rapid AI development, including a 'pre-crime' approach that predicts incidents and reports suspects to police before crimes occur.
- The firm's surveillance efforts contradict its self-image as a responsible alternative to OpenAI, especially after earlier refusing military contracts for mass surveillance but now hiring for national security sales.
- Anthropic uses Samdesk for real-time protest tracking; an executive was rerouted from a planned protest after 60 minutes' advanced notice via Samdesk.
- Anthropic routinely reports threats to police departments, but in one case refused to provide actual chat logs when reporting a user for mentioning an AR-15 and targeting the CEO, citing internal policy.
- The security team aims for predictive and preventative threat management, as stated by the security program manager; a job posting for an enterprise intelligence specialist ($180k-$230k) includes tracking activism, terrorism, and nation-state targeting.
- Anthropic's intelligence expansion aligns with efforts to designate AI as critical infrastructure, which would grant more access to federal intelligence and embolden companies to label civic engagement as extremism.
- CEO Dario Amodei acknowledges a public trust crisis regarding AI but continues the monitoring approach; security guards at Anthropic and OpenAI authorized a strike over low wages, leading Anthropic to advise work-from-home.