The U.S. AI Safety Institute, part of NIST, has partnered with Anthropic and OpenAI to enhance AI safety through collaborative research and model evaluation. These agreements aim to address safety risks and promote responsible AI development, marking a significant milestone in AI governance.