AI Safety Policy Entrepreneurship Fellowship 2026–2027: $5,000 Stipend, Policy Training and Frontier AI Governance Experience
The AI Safety Policy Entrepreneurship Fellowship 2026–2027 offers early- to mid-career professionals an opportunity to turn ideas about frontier artificial intelligence safety into practical policy solutions. Running from September 30, 2026, through February 28, 2027, the program provides a $5,000 stipend and policy training.
Overview
The AI Safety Policy Entrepreneurship Fellowship (PEF) 2026–2027 is a part-time program designed for individuals seeking to transition ideas regarding artificial intelligence safety, security, and governance into actionable policy measures. Rather than focusing solely on academic research, the fellowship emphasizes policy entrepreneurship, enabling participants to work part-time as affiliates of the Federation of American Scientists (FAS).
Selected participants engage in policy entrepreneurship training, group check-ins, reading groups, and monthly individual mentorship with dedicated FAS staff. The program also features two mandatory in-person gatherings:
- A California retreat tentatively scheduled for November 4–7, 2026.
- A capstone event in Washington, DC, tentatively scheduled for the week of February 22, 2027.
Benefits
- A $5,000 fellowship stipend for participation during the program period.
- An optional $1,000 merit award for the fellow who develops the most specific and actionable policy memo and advocacy plan.
- An additional travel stipend covering selected expenses for the California retreat and the Washington, DC capstone event.
- Professional editing support for written products developed throughout the fellowship.
- Access to FAS programming, curated updates on federal science and technology policy, and connections with relevant experts and policy actors.
Eligibility
The fellowship is best suited for early- to mid-career professionals possessing strong subject-matter expertise who want to gain practical public policy experience. Potential applicant backgrounds include:
- Technical artificial intelligence research
- Academia and think tanks
- Civil society and the technology industry
- Law, cybersecurity, and national security
- Public policy and science and technology policy disciplines
How to Apply
Candidates must submit their applications through the designated online platform. Applicants are required to create an account on the platform before submitting their materials and should avoid waiting until the final hours to begin the application process. Proposals must focus on a specific frontier AI safety challenge, present an actionable solution, and outline a practical strategy for real-world adoption.
Deadline
Applications must be submitted by 11:59 PM ET on September 7, 2026.