Anthropic argues that increasingly capable AI models, including Claude Mythos, heighten risks of misuse, driving a push for stronger safety measures, transparency, and regulatory readiness, and explains that donations are meant to elevate AI governance debates among policymakers.

The fundraising move comes as AI regulation debates intensify, with Wall Street Journal reporting highlighting tensions between regulation-minded groups and industry-friendly positions that could shape White House interactions.

Public First Action, launched by two former members of Congress, aims to influence AI policy through issue advertisements and is positioned as a counterweight to Leading the Future, a group backed by industry figures like Greg Brockman and Marc Andreessen.

Public First Action operates as a tax-exempt entity that runs issue-focused ads rather than candidate-specific messages, and it was created to counter stricter regulation efforts from rivals in the AI policy space.

Anthropic’s moves reflect a broader trend of tech firms engaging in public policy on AI regulation, national security, privacy, and economic impact as policymakers balance innovation with safety.

Regulatory context includes state-level efforts to police AI under existing consumer protection, privacy, and licensing laws, signaling reliance on current frameworks alongside new legislation.

The donation aligns with industry-wide efforts to shape U.S. policy amid rising calls for oversight, with midterm elections heating up and AI-related laws under consideration in several states.

The broader electoral backdrop features increased fundraising activity from Anthropic and peers, signaling heightened political engagement around AI safety and regulation ahead of the midterms.

While advanced AI promises benefits for science and economy, experts warn that without safeguards, critical infrastructure and security could be at risk.

Anthropic’s leadership, including CEO Dario Amodei, has advocated for stronger AI regulation with proposals for government oversight, testing, and the ability to block dangerous deployments if safety standards aren’t met.

Anthropic calls for tighter export controls on advanced chips and semiconductors and policies to prevent distillation or unauthorized copying of U.S. AI models for national security reasons.

Anthropic commits $20 million to Public First Action to advance AI regulation, emphasizing policy education and responsible governance rather than election campaigns, while Public First Action’s PACs have supported a mix of federal candidates.