GovAI DC Winter Fellowship 2027: A $21,000 Three-Month, In-Person Research Program in Washington for Emerging AI Governance and Policy Talent
GovAI’s DC Winter Fellowship 2027 was a full-time, in-person, three-month program in Washington, DC that paid a $21,000 stipend plus travel support for people building careers in American AI governance and policy.
GovAI DC Winter Fellowship 2027: A $21,000 Three-Month, In-Person Research Program in Washington for Emerging AI Governance and Policy Talent
The DC Winter Fellowship 2027 is now a historical reference. The Centre for the Governance of AI (GovAI) says on the official program page that applications for the DC Fellowship are closed. The page does not announce a replacement cycle or a new application deadline. The archive metadata retains the closed cycle’s original deadline, while this page makes clear that readers cannot apply to that round now.
The program itself was a three-month, full-time, in-person fellowship in Washington, DC. GovAI described it as a bipartisan opportunity for people who want to accelerate or launch a career in American AI governance and policy. Fellows were scheduled to join from January 18 to April 9, 2027, conduct a research project of their choice with expert guidance, and take part in workshops, seminars, peer feedback, and networking with people working on US AI policy.
That distinction between the program details and the application status matters. The stipend, format, eligibility rules, and selection stages below describe the 2027 cycle published by GovAI. They are useful for understanding the opportunity and preparing for a possible future announcement, but they are not an invitation to submit an application today.
Key details
| Item | Verified detail |
|---|---|
| Program | DC Winter Fellowship 2027 |
| Organisation | Centre for the Governance of AI (GovAI) |
| Status | Applications closed; historical reference |
| Format | Full-time and in person in Washington, DC |
| Fellowship term | January 18 to April 9, 2027 |
| Funding | $21,000 stipend plus support for travel to DC |
| Workplace support | Weekday lunch and a desk in GovAI’s DC office |
| Work authorization | Right to work in the United States required; no visa sponsorship |
| Degree requirement | No strict degree requirement |
| Part-time possibility | Possible for some DC-based researchers, government staffers, or relevant professionals, at a pro-rated salary |
| Official page | https://www.governance.ai/post/dc-winter-fellowship-2027 |
| Application questions | [email protected] |
What the fellowship was designed to provide
GovAI’s DC fellowship combined a paid placement with a guided research project. The central assignment was not a pre-set internship task. Each fellow chose a topic, spent the first two weeks exploring the AI policy field, and then settled on a proposal with input from a supervisor. A fellow’s work could become a report, white paper, journal article, op-ed, or blog post aimed at an audience relevant to AI policy.
The structure was intended to help participants learn how policy questions are selected and developed, not just how to produce a final document. A supervisor from the GovAI team or network provided expert guidance, and weekly meetings gave the fellow a regular place to test assumptions, narrow the question, and decide what evidence would support a useful output. The Fellowship Manager and wider team were also available to help identify a project that was both feasible within the term and valuable to people making or studying policy.
Research was only one part of the experience. GovAI planned Q&A sessions with AI policy experts, hands-on workshops and seminars for building relevant skills and subject knowledge, work-in-progress meetings for peer feedback, and social events. Fellows were encouraged to discuss follow-on career options with GovAI’s staff and network. The program therefore suited someone who wanted both a substantial piece of work and a closer view of the professional community around US AI governance.
Funding, location, and working arrangement
The published package was a $21,000 stipend for the three-month term, together with support for traveling to Washington, DC. GovAI stated that the role was in person and that fellows would have a desk in its DC office. Weekday lunch was also provided. These details make the opportunity different from a remote writing fellowship or a short unpaid policy internship: participants were expected to live and work in Washington for the term and to take part in the office-based program.
The standard arrangement was full time. GovAI also said it might facilitate a part-time fellowship at a pro-rated salary for a current researcher at a DC think tank, a staffer in a government office, or another DC-based professional doing relevant AI governance work. That was presented as an accommodation for particular circumstances, not as a general remote option. The official page also said GovAI might make arrangements if the pay created a barrier to participation and directed candidates with questions about their requirements to contact Martin Fukui with a CV and a brief explanation.
Who the 2027 cycle sought
GovAI invited applicants from government, academia, industry, and civil society. It described the target group broadly: early-career people interested in entering US AI policy as well as experienced professionals moving into emerging technology policy from an adjacent field. Promising undergraduates could apply, and the page explicitly said there were no specific role requirements and no strict degree requirements.
Relevant preparation could come from public policy, political science, engineering, computer science, economics, biosecurity, cybersecurity, China studies, or risk management such as assurance and auditing. That list was illustrative rather than exhaustive. The better question for a prospective applicant is whether their experience, research, or graduate study would help them make a careful contribution to a US AI policy question.
Two practical conditions were firm. Applicants needed the right to work in the United States, because GovAI did not sponsor visas for this fellowship. They also needed to be able to participate in a full-time, in-person program in Washington, subject to the limited part-time possibility described above. Someone who could not relocate or who needed visa sponsorship would not meet the published conditions for this cycle.
The program also placed unusual emphasis on intellectual diversity and bipartisan engagement. GovAI said it wanted fellows with a range of perspectives and political backgrounds who could work in good faith on shared AI policy challenges. This does not mean an applicant had to avoid having views. It means the application needed to show that they could distinguish analysis from advocacy, express uncertainty, respond to criticism, and work with people who did not share their assumptions.
Selection criteria
The official criteria focused on four areas:
- Quality of work: the ability to produce clear, insightful, and even-handed research or policy advice, with especially strong reasoning and concise writing.
- Relevant expertise: skills or knowledge that could help with topics relevant to AI policy. GovAI noted that expertise can take many forms.
- Judgment: the ability to prioritize among possible research or policy directions and assess whether a direction is feasible.
- Team fit: openness to feedback, intellectual honesty and rigor, comfort expressing uncertainty, and a serious interest in using one’s career to help navigate the transition to a world with advanced AI.
These criteria point toward an application built around evidence rather than broad enthusiasm. A good candidate would explain a question they care about, show how they have handled difficult material before, and demonstrate that they can change course when the evidence or feedback warrants it. A polished CV alone would not establish those qualities. Clear writing and a grounded account of past work would do more.
How the application process worked
The 2027 process had three stages, but the current status is closed and GovAI is not accepting applications for this cycle.
- Written application and abstract reasoning assessment. Candidates submitted a written application. Those selected for the next step completed a short automated abstract reasoning assessment of approximately 20 minutes. The assessment served as a route into the second round, rather than an extra application that every candidate necessarily completed.
- Paid remote work test. Candidates invited to the second round completed a paid remote work test. GovAI said it expected to invite candidates to this stage in August. The test was meant to show how a candidate approaches and communicates a governance problem in practice.
- Remote interview and reference checks. The third stage took place in September and consisted of a remote interview and reference checks. GovAI expected to communicate final decisions in October.
For a future cycle, applicants should treat the work test as a real selection stage, not a formality. They would need to read the prompt carefully, state assumptions, separate known facts from speculation, and deliver a conclusion that a policy audience could use. References should also be lined up early enough for the final stage. None of these steps is currently available for the closed 2027 round.
Research topics and outputs
The fellowship did not require applicants to arrive with a final research proposal. GovAI’s model gave the first two weeks to exploring the field before the fellow settled on a project with their supervisor. That approach favored applicants who could start with a meaningful question while remaining willing to narrow or revise it.
A useful project would connect an important AI governance problem to a tractable method and a clear audience. Depending on the question, the output might be a report for policymakers, a white paper, a journal article, an op-ed, or a blog post. The official description did not promise publication, a particular journal, or a specific policy placement. It promised the chance to do independent research with expert guidance and to choose an output appropriate to the work.
The program’s bipartisan focus also shaped the expected standard. A strong project would be even-handed, explicit about uncertainty, attentive to how US political institutions work, and practical about what decision-makers could actually do. That standard applies whether the topic is technical governance, national security, economics, regulation, or another area relevant to American AI policy.
What readers should do now
There is no current application step for the DC Winter Fellowship 2027. Readers should not send a new application to the closed round or treat the archived deadline as an active date. The official program page is the place to check if GovAI announces another DC fellowship or changes the status of this listing.
Someone preparing for a possible future cycle can still use the published requirements productively. First, confirm the practical constraints: US work authorization, ability to be in Washington, and enough availability for a full-time term unless a specific part-time arrangement is discussed. Second, read GovAI’s research and policy work and identify one or two questions that matter to US AI governance. Third, collect examples of clear writing, research, policy analysis, or other work that shows reasoning and judgment. Finally, ask potential references well in advance and be ready to complete a timed reasoning exercise and paid work test if a new process opens.
Frequently asked questions
Can I apply to the 2027 DC Winter Fellowship now? No. GovAI’s official page says applications for the DC Fellowship are closed. This page is an archive entry for the closed cycle.
Was the program remote? No. GovAI described it as an in-person role in Washington, DC. The normal arrangement was full time, with a desk in the DC office and weekday lunch. A limited part-time arrangement might have been available for certain DC-based professionals.
How much was the stipend? The published stipend was $21,000, plus support for traveling to Washington, DC. GovAI also said it might make accommodations if pay was a barrier to participation.
Did applicants need a particular degree? No. GovAI said there were no strict degree requirements and welcomed promising undergraduates as well as people with professional or research experience.
Could someone outside the United States apply? Only if they already had the right to work in the United States. GovAI said it did not sponsor visas for this fellowship.
What would a fellow research? The fellow chose a topic relevant to AI policy, explored the field during the first two weeks, and developed a proposal with a supervisor. Possible outputs included a report, white paper, journal article, op-ed, or blog post.
Has GovAI announced the next cycle? The official DC Winter Fellowship page says the 2027 applications are closed and does not announce a new cycle or deadline. Check GovAI’s opportunities page for any later announcement.
Official source
The verified source for this archive entry is GovAI’s official announcement:
- Program page: https://www.governance.ai/post/dc-winter-fellowship-2027
- Organisation: Centre for the Governance of AI (GovAI)
- Application questions: [email protected]
- Special-circumstances contact named by GovAI: [email protected]
The program page is still available, but its status is closed. The 2027 opportunity should therefore be read as a historical reference until GovAI publishes a new announcement with current dates and application instructions.