Anthropic, the company behind Claude, has shared a draft federal Bill with members of Congress that would set national safety and security rules for the most powerful frontier AI models. The Bill materializes the ideas in Anthropic's "Advanced AI Framework," published in June 2026, which aims to guard against the most serious AI risks without stifling innovation. In that framework, the rules would apply only to the largest AI developers, i.e., those that train very large models and either earn more than $500 million a year from AI or spend more than $1 billion a year on AI research.