Anthropic's Clark calls for structured AI testing and information sharing

Clark: Jack Clark is a co-founder of Anthropic and a leading commentator on AI policy, safety, and risks. He is directly quoted in the news calling for AI testing and information sharing mechanisms that do not depend solely on voluntary commitments from companies. Anthropic: Anthropic is an AI research company known for developing the Claude series of large language models with a strong emphasis on safety and alignment. In this news, the company is represented by co-founder Jack Clark advocating for more structured approaches to AI model testing and information sharing beyond voluntary industry efforts. Safety Collaboration: Anthropic recently announced expanded partnerships with independent evaluators including METR, granting them deep access to models and data for risk assessments. Transparency Practices: The company has begun publishing standalone reports on model behaviors observed during evaluations, separate from standard system cards.