Institute for Law & AI Summer Research Fellowship (US): 2026 Cycle Closed, 2027 Not Announced

The 2026 round is closed. The Institute for Law & AI (LawAI) lists its US-track summer fellowship at law-ai.org/srf-us under the heading “[Closed] Summer Research Fellowship (US)”. Applications closed on February 7, 2026, at 11:59 pm ET, and the page says late applications could not be assessed. The official seasonal fellowship hub lists the current US summer fellowship as closed and does not announce a 2027 call. This page is therefore a historical reference to the verified 2026 cycle, not an invitation to apply.

The other seasonal tracks shown on LawAI’s programme hub are also marked closed. The hub is the official place to check for a future announcement, because it collects the summer and winter US, EU Law and Legal Frontiers fellowships in one place.

This page previously pointed at a raw Airtable application form. That form belonged to the closed round and is no longer the right destination; the official programme pages above are.

At a Glance

What the fellowship actually is

LawAI runs seasonal research fellowships for people who can think like lawyers and want to apply that skill to artificial intelligence. The US track is the one this page covers, and its subject matter is American: how existing statutes, agencies and doctrines handle AI systems, and what should change.

The work is genuine research rather than administrative support. Fellows are paired with LawAI research staff and affiliates and produce written output — memos, research notes, drafts that can feed into published work. The questions on the table are the ones being argued about in Washington right now: liability allocation when an AI system causes harm, how export controls apply to model weights and compute, whether civil rights statutes reach algorithmic decision-making in housing and hiring, what agency authority already exists versus what would need new legislation.

Three things make the programme unusual among law-adjacent summer positions. It pays properly. It is remote-first, so you are not forced to relocate for ten weeks. And it treats fellows as colleagues rather than as a labour pool, which shows up in the mentorship structure and in where alumni end up — past fellows have moved into roles in federal government, at AI developers, in universities, at think tanks, and at LawAI itself.

The money

For the US track, the weekly subsidy is $1,500, and LawAI states that candidates with exceptional experience or qualifications may be eligible for more. Across a ten-week fellowship that is $15,000 before any expense coverage.

On top of the weekly rate, LawAI covers flights, accommodation and meals for the in-person week in Washington, D.C. or Berkeley. Fellows also attend the Institute’s Summer Institute on Law and AI in Washington, D.C., with travel and lodging covered.

The comparison worth making is not against a big-firm summer associate salary, which is a different employment category, but against the unpaid or nominally paid policy internships that are a common route into this field. The official terms make this a paid research placement rather than an unpaid volunteer role.

The amount above applies to this US Law and Policy track. Other LawAI tracks have separate terms, so do not use the US figure as a shortcut for the EU Law or Legal Frontiers fellowships.

Who it is for

Prior AI-specific research experience is not required. Prior legal grounding is. LawAI asks that applicants for the US and EU Law tracks have prior legal experience through study or work in the relevant jurisdiction, which in practice means US law school, US practice, or academic work anchored in American law.

The profile that fits:

• A second- or third-year US law student who has taken administrative law, has written something substantial on technology or discrimination, and wants to test policy work against the firm track.
• A practising lawyer in technology, civil rights, national security or regulatory work who wants a concentrated stretch on AI questions and a network outside their current employer.
• A doctoral student or academic working at the seam of law, technology and public policy who wants structure and colleagues for an AI-focused project.

You do not need to write code or understand transformer internals. You do need to reason clearly about what these systems do in the world and where legal instruments bite.

The commitment is real: full-time for the ten-week period, plus attendance at the in-person week. LawAI has historically considered modest adjustments to start and end dates but prefers full participation. Work authorisation for wherever you will be based is a hard requirement — the Institute has said it will make reasonable efforts to support visa processes for candidates who receive offers, but the outcome rests with immigration authorities, not with LawAI.

How the selection process ran

The 2026 round used a four-step process. It is useful as a record of what applicants were asked to do, but a future call could change the sequence or materials:

1. An online application form. The official page estimated roughly twenty minutes. It required submission by 11:59 pm ET on February 7, 2026, and the page states that late applications would not be assessed.

2. An asynchronous exploratory interview, around thirty minutes, in which candidates record answers to set questions about AI and its legal and social consequences.

3. A research interview, around sixty minutes, with a member of the research staff, covering your research interests, goals and fit for the programme.

4. A decision. LawAI said it aimed to provide all applicants with a final decision by mid-March and invited applicants to note a required decision date on the form.

Treat that structure as the planning baseline for a future round while checking the official page for changes when a call actually opens.

How to be ready for the next call

There is no announced next US-track round. Do not infer an opening month or invent a future deadline from earlier cycles. Use the official seasonal fellowship hub and LawAI’s newsletter for a confirmed announcement.

Narrow your interest to something researchable. “AI and privacy” is not a research interest; “whether the FCRA’s adverse-action notice requirements can carry the weight being asked of them when scoring models are opaque” is. Name one or two questions you would want to work on. You are not committing to them — you are demonstrating that you can take a sprawling topic and cut a tractable question out of it.

Write something now. If everything you have written is locked behind a firm’s or a court’s confidentiality, you have no writing sample. Draft a five-page memo or a blog-length piece on a live AI law question, hand it to a professor or a colleague, and let them cut it apart. A piece written deliberately for this audience beats a seminar paper you happened to have.

Read what LawAI publishes. The Institute puts out research and commentary, and the fellowship page points applicants to its United States Law and Policy Workstream and Research Directions. Reviewing those materials helps you identify questions that actually fit the programme’s work.

Practise the asynchronous interview format. Recording yourself answering an open question in two or three minutes is a specific skill, and most people are bad at it the first time. Structure each answer: state the problem concretely, identify which legal instruments plausibly apply, then be honest about the trade-offs and what you are uncertain about. Overconfidence reads worse than calibrated doubt to this audience.

Sort your logistics early. If your availability is complicated or your immigration status is unsettled, work out the facts before you apply and state them plainly on the form. Concealing a constraint until after an offer creates exactly the kind of friction that makes a candidate expensive.

Check the other seasonal tracks only when they are announced. The hub includes Winter Research Fellowship pages as well as the summer tracks, but the current listings are closed. Their dates, eligibility and materials are separate from the US summer fellowship and should be verified on their own official pages.

Frequently asked questions

Is anything open right now?

No. The official seasonal fellowship hub marks the listed Summer US, Summer EU Law, Summer Legal Frontiers and Winter fellowship pages as closed.

When exactly did applications close?

The official page gives the closing time as 11:59 pm ET on February 7, 2026. It also says that the application took approximately twenty minutes and that late applications could not be assessed.

Will there be a 2027 round?

Nothing for 2027 is announced on the US fellowship page or the seasonal hub. Watch the programme hub and LawAI’s newsletter for a confirmed announcement rather than relying on the 2026 schedule.

Do I have to be a US citizen?

Nationality is not the stated filter; prior legal study or work in the US is, along with valid work authorisation for wherever you will be based during the fellowship.

Can I do it part-time?

The fellowship is designed as full-time work for the ten-week period. LawAI has considered adjustments to start and end dates in exceptional cases, but it is not structured to sit alongside another demanding role.

Does the fellowship lead anywhere?

Often. LawAI points to continuing collaboration with former fellows through follow-on funding, event invitations and, in some cases, roles at the Institute. Alumni have gone on to federal government, AI developers, academia and policy organisations.

What to do now

Do not submit anything to the old Airtable form; that round is finished. Instead:

1. Open law-ai.org/seasonal-research-fellowships and read how the tracks differ, so you know which one fits your background if a new call appears.
2. Sign up for the Institute’s newsletter and check its official updates for a future announcement.
3. Read the United States Law and Policy Workstream and Research Directions linked from the fellowship page, then prepare a concise research idea and writing sample for a future application.

The applicants who do well in this process are rarely the ones who found the listing a week before it closed. They are the ones who had already decided what they wanted to work on.

Official programme page: Summer Research Fellowship (US) — Institute for Law & AI