Regulatory Consulting at Federation of American Scientists (FAS)

Overview

The Federation of American Scientists is a nonpartisan science-and-technology policy organization with unusual moral lineage: founded in November 1945 as the Federation of Atomic Scientists by researchers who built the first atomic bombs during the Manhattan Project, on the conviction that technically trained people have an ethical obligation to ensure their work benefits humankind. That founding logic maps directly onto today’s AI debates, and FAS has become a significant voice on AI safety, security, and government capacity. Under CEO Daniel Correa — who previously founded the Day One Project and led Stanford’s Technology and Public Policy Project — the organization has broadened from nuclear-weapons transparency into the full span of science policymaking, with a consistent emphasis on getting scientists and technologists into government service. Its Day One Project is the engine: since 2020 it has helped a growing community of contributors develop actionable policy ideas, a striking number of which have become actual policy. On AI specifically, FAS has published detailed recommendations structured around unleashing AI innovation, accelerating adoption, ensuring secure and trustworthy AI, and strengthening U.S. government institutions — including proposals for a federal center of excellence to expand state and local capacity for AI procurement and use.

AI-Focused Regulatory Work

FAS’s defining AI project is an 18-month collaboration with the Future of Life Institute that concluded in May 2026 with the Converging Risks report, presented alongside the AI Impact Awards at the International Spy Museum in Washington. The partnership built a series of convenings and reports across the intersections of AI with biosecurity, cybersecurity, nuclear command and control, military integration, and frontier AI governance — an agenda that only an organization with FAS’s nuclear-era credibility could credibly assemble. CEO Daniel Correa framed the stakes directly: AI is no longer a single-domain challenge, and FLI’s Anthony Aguirre called it a force multiplier reshaping the risk landscape across nuclear, biological, cyber, and military systems simultaneously, faster than institutions can adapt. FAS’s second mode of influence is the Day One Project’s steady supply of concrete, implementable ideas: a federal center of excellence for state and local AI procurement, ‘fuzzy provenance’ solutions at NIST for authenticating AI-generated text, AI provenance click-throughs, and a Disaster Management and Technology center of excellence. When the administration sought public comment on its AI Action Plan in February 2025, FAS responded with ‘Securing American AI Leadership,’ a four-pillar strategic plan developed with AI experts, industry leaders, and research institutions — then published a candid follow-up assessing what the final Action Plan got right and what it missed. That combination of convening power, an idea pipeline into government, and a willingness to grade the results makes FAS an entrepreneurial and unusually practical player in AI policy.

Notable AI Policy Work

• Concluded an 18-month partnership with the Future of Life Institute producing the Converging Risks report (May 2026) on AI’s intersections with biosecurity, cybersecurity, nuclear command and control, military integration, and frontier AI governance.
• Runs the Day One Project, which since 2020 has developed actionable science-and-technology policy ideas — many of which have become policy — including proposals for a federal center of excellence to expand state and local AI procurement capacity.
• Published ‘Securing American AI Leadership: A Strategic Action Plan for Innovation, Adoption, and Trust’ as its response to the White House AI Action Plan RFI, built on four pillars with input from AI experts and research institutions.
• Published a candid assessment of ‘What’s Progress and What’s Not in the AI Action Plan’ after the administration released its final plan (July 2025).
• Developed AI-provenance and information-integrity proposals — including ‘fuzzy provenance’ solutions for NIST to authenticate AI-generated text and AI provenance click-throughs — plus a Disaster Management and Technology (DMAT) Center of Excellence concept.
• Presents the AI Impact Awards recognizing civil-society and other contributions to AI policy (2026 honorees included Encode founder Sneha Revanur).

Thought Leadership

1. Converging Risks: AI and the Future of Global Security (May 2026)
   The capstone of an 18-month FAS–Future of Life Institute project mapping AI’s compounding risks across biosecurity, cyber, nuclear command and control, and military systems.
2. Securing American AI Leadership: A Strategic Action Plan for Innovation, Adoption, and Trust
   FAS’s four-pillar response to the White House AI Action Plan RFI, developed with AI experts and research institutions.
3. What’s Progress and What’s Not in the AI Action Plan
   FAS’s candid assessment of how the administration’s final AI Action Plan measured up against its recommendations.
4. Day One Project
   FAS’s idea pipeline democratizing policymaking — developing science and technology proposals, many of which have become actual policy.
5. Bold Emerging Technologies and AI Policy for 2025 and Beyond
   Day One Project proposals on AI safety, provenance, disaster response, and federal AI investment.
6. FAS & FLI Present Converging Risks Report and AI Impact Awards
   Coverage of the May 2026 gala concluding the FAS–FLI project, with leadership framing AI as a cross-domain force multiplier outpacing institutional adaptation.

Clients

Nonpartisan nonprofit; no commercial clients. Engages the White House Office of Science and Technology Policy, NIST, Congress, and federal agencies through public comment and the Day One Project. Key partner: the Future of Life Institute (Converging Risks). Community contributors span academia, industry, and research institutions.