Applications are now open for the AI Safety Policy Entrepreneurship Fellowship, a programme supporting individuals with practical ideas for tackling frontier AI safety, security and governance challenges.

The fellowship is intended to help participants develop policy ideas from the concept stage towards real-world implementation. Fellows will work on engaging decision-makers, forming coalitions and creating practical routes for policy adoption.

Selected participants will receive a $5,000 stipend, while the fellow who produces the most specific and actionable policy memo and advocacy plan will be eligible for an additional $1,000 merit award.

The programme will also provide travel support for selected expenses, policy entrepreneurship training, professional editing assistance for written work and monthly one-on-one check-ins with an FAS staff mentor.

Fellows will receive support in connecting with relevant experts and stakeholders, as well as invitations to FAS programmes and policy-focused sessions.

As part of the fellowship, participants will identify a specific frontier AI safety challenge and develop a policy memo that outlines a practical solution and a strategy for securing its adoption.

The programme will include training sessions and reading groups, with fellows learning how to identify policy opportunities, build coalitions and engage decision-makers.

Participants will also examine possible implementation routes through legislation, regulation, standards, procurement and other governance mechanisms. The programme includes a retreat in California and a capstone event in Washington, DC.

Applicants are encouraged to submit proposals addressing areas such as technical AI safety, AI governance and policy, AI risk assessment and mitigation, transparency and accountability in AI, cybersecurity, AI capability thresholds, risk taxonomies, critical infrastructure, compute governance, AI model proliferation and governance of catastrophic AI risks.

The fellowship is open to applicants from a range of professional backgrounds and is particularly suited to early- and mid-career professionals with limited direct public policy experience who want to translate their technical or subject-matter expertise into practical policy impact.

Potential applicants may come from fields including AI research, academia, think tanks, civil society, industry, law, cybersecurity, national security and other related areas.

The fellowship will run from September 30, 2026, to February 28, 2027. It is a part-time and hybrid programme requiring approximately five hours per week.

Participants will also be required to attend twice-monthly group check-ins and other mandatory programme activities.

The California retreat is tentatively scheduled for November 4–7, 2026, while the Washington, DC capstone is tentatively planned for the week of February 22, 2027.

Applications must be submitted by September 7, 2026, at 11:59 PM ET.

The fellowship is aimed particularly at professionals who already understand a significant AI-related challenge and have an idea for how governments, institutions or other decision-makers could address it.

Rather than following the model of a traditional research fellowship, the programme places a strong emphasis on implementation, helping participants develop ideas that could potentially contribute to real-world policy and governance.

Interested applicants can review the full requirements and submit their applications through the official FAS fellowship page.

Source: msme Africa online