[UPDATE] Anthropic red team says small AI models may be months from self-replicating on IoT devices
Anthropic's Frontier Red Team published research saying small models 'may be months away' from being able to copy themselves onto poorly secured consumer IoT devices. The researchers did not show a fully autonomous worm on real hardware, and the setups they tested were slow, taking days per copy. BleepingComputer and Al Jazeera both describe the finding as a trajectory warning, not a current threat. Still to come: responses from device makers and network providers on the detection and quarantine steps the report proposes.