Policy Development, Technical AI Governance

About the role

Anthropic believes AI capabilities are improving rapidly, and that AI safety policy needs to keep pace. Our Policy Development team translates technical insights into workable new policy ideas for governing frontier AI. As a policy analyst, you will drive our work on technical and cross-lab safety practices, deepen our work on frontier AI legislation, and develop new technical governance ideas that you pressure-test with researchers and carry into policy proposals. You will work closely with policy colleagues and with research, alignment, safety, and legal teams, and you’ll move easily between technical detail and how policy actually gets made.

Key responsibilities

• Help develop technical safety best practices and standards for frontier development, working closely with policy, research, and safeguards teams
• Support cross-lab discussions and initiatives and serve as the Anthropic representative in key standards initiatives and groups
• Review and give substantive input on federal, state, and global frontier AI safety bills, including on risk reporting, independent evaluation, and incident reporting, helping shape AI safety legislation development
• Work with our external affairs policy teams to strengthen our frontier AI legislative framework
• Contribute research and analysis across the policy team’s pacing workstreams, including developing novel policy proposals that address challenges related to the acceleration of AI R&D
• Engage with the external AI safety and governance community to track new thinking and feed it into our own
• Write clear memos and briefings that help leadership align on decisions

Minimum qualifications

• Are fluent in frontier AI safety and technical AI governance: the main players, proposals, and debates
• Can read legislative text and comment usefully on it
• Have expertise in any of the following domains: loss-of-control, alignment, compute governance, independent evaluation, model weight security, risk reporting
• Write clearly for researchers, leadership, and policymakers
• Can pick up new technical topics quickly and translate them to the broader policy landscape
• Are comfortable with ambiguity and can own a project end to end

Preferred qualifications

• Are excited about standards work, and also want to contribute across the wider technical governance agenda
• Have direct experience in an AI governance or safety research organization or frontier lab
• Have experience working closely with standards bodies or industry consortia
• Have experience developing and shaping technical legislation
• Have published novel policy thinking on AI governance