This panel will explore how the AI community can leverage an open science approach to examine new use cases and deployments for AI, develop and improve new metrics and benchmarks, identify issues for remediation, and generate new mitigations. The panel will also assess how an open science approach that leverages researchers’ direct access to model weights and architectures can manage and alleviate concerns about AI risk areas while pushing the industry forward and creating economic opportunity.