Mandate AI Safety Regulations Before Capabilities Outpace Control

Pass binding AI safety legislation now, before the technology outpaces our ability to govern it. The most powerful AI systems are being developed and deployed without mandatory safety evaluations, liability frameworks, or capability limits. That is not innovation — it is negligence at civilizational scale. Other critical technologies — pharmaceuticals, aviation, nuclear energy — require rigorous pre-deployment testing and ongoing oversight. AI, which poses risks that experts across the political spectrum describe as potentially irreversible, gets none of that. Voluntary commitments from the companies profiting from rapid deployment are not a substitute for law. Mandate independent safety audits, establish hard capability thresholds that trigger regulatory review, and create real accountability when systems cause harm. The window to act before these systems become ungovernable is closing fast.