1. United States
2. Ore.
3. Letter

Investigate and pass emergency guardrails on AI development before self-replicating AI worms become an uncontrollable threat. Anthropic's own researchers warned this week that autonomous AI malware capable of spreading through smart home devices, security cameras, and video-conferencing units could be "months away." Their report describes a worst-case scenario where multiple AI worms "swarm" together to attack datacenters, with no centralized point to shut them off. This is not a hypothetical from outside critics. It comes from Anthropic's frontier red team, the unit built specifically to stress-test their own systems. Even Anthropic's CEO, Dario Amodei, and OpenAI's CEO, Sam Altman, have called for a slowdown. Yet both companies have continued releasing new models anyway. That gap between what these executives say and what they do is exactly why legislation cannot wait for industry self-regulation. Homes and small businesses would be the first targets. Ordinary people with smart ovens and automatic lawnmowers should not bear the risk of an industry that is moving faster than its own safety teams can manage. Pass binding oversight now.