AI Worms Are Coming for Your Smart TV - Anthropic's Self-Replicating AI Research

Anthropic's Frontier Red Team published research showing small AI models may be months away from self-replicating across cheap IoT devices - no datacenter needed. They ran Gemma4-26B-A4B distributed across 15 off-brand Android TV boxes (~$55 each), networked together, running inference at 0.39 tokens/sec. A 9B model got inference working on 8 of 9 tested hardware types on its own. Models like GLM-5.2 can already complete the full replication chain end-to-end in simplified tests. The key risk: such a "worm" has no centralized shutdown point, unlike a datacenter-based attack - though it would be slow, with a doubling time of days. Source: https://www.anthropic.com/research/se... Also covered by CNBC, Bleeping Computer, ABC News, and Fortune. #AI #AISecurity #Anthropic #CyberSecurity #IoTSecurity #AIWorm #TechNews #AIEngineering #MachineLearning