Anthropic is donating another $20 million to Public First Action

Anthropic announced it is contributing an additional $20 million to Public First Action, bringing its total financial support for the nonpartisan organization to $40 million. Public First Action educates the public about artificial intelligence and works with lawmakers from across the political spectrum to establish sensible AI safeguards. According to Anthropic, the donations are made exclusively to support public education and the organization's policy mission and cannot be used to influence the election of any candidate for office. The first donation was made in February 2026 to promote policies maintaining safeguards, sustaining American AI leadership, and demanding transparency from builders of powerful models. Anthropic argues that as AI models improve rapidly, such as its Claude Mythos Preview discovering thousands of high-severity software vulnerabilities, transparency alone is insufficient. The company contends that policymakers must put forward measures to mitigate risks, including allowing governments to verify safety claims, enforce safe practices with civil penalties, and slow or block the deployment of models posing risks of catastrophic harm. Furthermore, Anthropic supports tightening export controls on advanced chips and curbing illicit model access to maintain American AI leadership. Beyond its donation, Anthropic is partnering with METR for safety evaluations, strengthening election safeguards ahead of US midterms, and opening a new office in Singapore alongside signing a memorandum of understanding with the country's government to advance safe AI adoption.