1315 ET - Policymakers need to get in the same room with AI company leaders this week to address the latter group's request for more industry oversight and slower development of the technology, Melius Research analysts say in a research note. Anthropic CEO Dario Amodei has called for national legislation requiring frontier-model testing, employee-level access for external evaluators and tighter controls for adversaries in China, they say. Some senators are drafting a bipartisan bill that could be introduced this week, which would likely try to impose a "duty of care" on frontier developers, the analysts say. That would give the government authority to block an unsafe model release and create a framework for labs to test for catastrophic cyber, biological and nuclear capability, they say.