Anthropic Donates Another $20 Million to Public First Action, Bringing Total to $40 Million
On July 21, 2026, Anthropic announced an additional $20 million donation to Public First Action, a nonpartisan nonprofit focused on AI education and policy. Combined with its initial February 2026 donation, Anthropic’s total support now totals $40 million.
Details
- Amount and timeline: Announced July 21, 2026; follows an initial $20 million donation in February 2026, bringing the cumulative total to $40 million
- Restrictions on use: Both donations are earmarked strictly for public education and policy purposes and explicitly cannot be used to influence federal, state, or local elections
- Rationale: Anthropic points to its newest model, Claude Mythos Preview, discovering “thousands of high-severity vulnerabilities” across major operating systems and browsers as evidence of how rapidly AI capabilities are advancing, and warns that such capabilities in the wrong hands could threaten critical infrastructure like hospitals and power grids
- Policy measures advocated: Government verification of safety claims, civil penalties for unsafe practices, independent model evaluation, mandatory security programs, export controls on advanced chips, and restrictions on illicit model access
- About Public First Action: A nonpartisan nonprofit that works with “Republicans, Democrats, and Independents” to advance sensible AI safeguards and educate the public about artificial intelligence
How to try it
- This isn’t a consumer product launch but a policy and advocacy funding announcement from Anthropic
- Full details on the policy rationale and the donation itself are available on Anthropic’s official announcement page