OpenAI and Anthropic Neared Deal to Stress Test Each Other’s AI

This collaborative effort marks a significant shift toward prioritizing safety through rigorous peer-reviewed stress testing of foundational models. By pooling resources and expertise, these industry rivals are establishing new standards for identifying vulnerabilities before public deployment.