Anthropic red team creates autonomous AI worm for smart TVs
October 19, 2026
Anthropic's security researchers successfully built a malicious program capable of autonomously spreading across 15 internet-connected smart TVs. The exercise demonstrates the potential for agentic AI to exploit network vulnerabilities and propagate through IoT ecosystems.
HOW THIS AFFECTS YOU
●
builderYou should prioritize hardening IoT device security against agent-driven exploits.
●
policyYou must prepare for new classes of autonomous network threats.