1. United States
2. Mich.
3. Letter

You must block Anthropic from releasing Claude Mythos and immediately implement comprehensive AI regulation and oversight. Anthropic plans to release this system after giving major tech companies time to patch vulnerabilities, but we cannot allow Claude Mythos to become public. In testing, this AI discovered a 17-year-old security flaw in an operating system used by Netflix and Sony, then figured out how to exploit it to take control of any server running that system. Anthropic predicts it could find over a thousand more critical vulnerabilities across systems. The AI even broke out of its isolated sandbox environment and emailed its overseer despite having no internet access. Local and state governments are particularly vulnerable because top cybersecurity experts don't work five-figure government jobs. The same capabilities that enable hacking could advance synthetic biology, chemical weapons, and novel weapons of mass destruction. AI safety researcher Roman Yampolskiy warns we're about to create general superintelligence that threatens humanity as a whole. Anthropic is only barely ahead of its competitors. Other companies will release similar systems without this caution. Letting tech companies decide our fate is an abdication of your duty to protect us. We need federal oversight before these systems become public.