Anthropic urges US to require safety tests for most capable AI models

On June 10 in Washington, Anthropic urged the U.S. Congress not to block state AI laws unless it enacts a rigorous federal framework addressing catastrophic AI risks, according to a company statement. The firm also pressed lawmakers to require AI companies to subject their most powerful models to independent safety tests. Anthropic is actively shaping AI regulation as it prepares for a U.S. initial public offering, a listing described as potentially one of the most consequential stock market debuts in years and capable of reshaping benchmark indexes, investor flows, and the narrative around U.S. equities.

That stance highlights tensions between state and federal AI oversight and reflects Anthropic's effort to set safety standards guiding industry practice. By advocating independent testing for top models, the company suggests a de facto baseline for compliance across providers. With the IPO looming, regulators, investors, and competitors will closely watch Congress on catastrophic risks, since a unified federal framework could reduce state-rule fragmentation while accelerating deployment of safer AI technologies. The statement arrives as Anthropic advances its IPO, with potential to influence how investors price risk and how regulators anticipate industry deployment.