Prudent Guardrails For The AI Race

New York Assemblymember Alex Bores introduced the RAISE Act, targeting only the largest AI developers spending over $100 million on model training. The legislation requires safety plans, incident disclosure, and prohibits releasing models that show unreasonable risk during testing, specifically focusing on catastrophic threats like bioweapons or infrastructure attacks. Unlike federal voluntary guidelines, this state-level approach provides enforceable standards with real consequences for frontier AI systems.