In this eye-opening experiment, cybersecurity creator Cyb3rMaddy tasks Claude (Anthropic’s AI coding agent) with turning a clean VM into a full telemetry surveillance system. The AI autonomously builds a FastAPI backend, deploys collectors for terminal commands, browser history, file activity, and Git changes, then delivers a slick cyberpunk dashboard that summarizes user behavior in real time — complete with suspicious activity alerts. While the system excels at operational pattern analysis and timeline reconstruction, Claude refuses to cross into personality or psychological profiling, highlighting both the alarming ease of AI-powered workplace (or personal) monitoring and the remaining guardrails that still require human judgment. A stark look at how quickly “AI collaborator” becomes “AI overseer.”
