Do you want to help make powerful AI safe, whatever your degree? A 15-week, paid fellowship in London is open to people from many fields. Applications close on 1 November 2026.
Pivotal Research is taking applications for its AI Safety Research Fellowship, Q1 2027 cohort. It runs for 15 weeks, full-time, from 18 January to 30 April 2027. Fellows work in person at LISA, the London Initiative for Safe AI. The stipend is £15,000, or £18,000 for Senior Fellows. Pivotal also pays for travel to London and gives £1,800 a month toward housing to fellows who live elsewhere. The deadline is 23:59 Anywhere on Earth on Sunday 1 November, six days from now.
Who can apply? Anyone aged 18 or older who wants to work on AI safety or AI governance. Pivotal says background matters less than motivation and ability. Past fellows have come from machine learning, philosophy, policy, physics and biology. They range from first-year students to PhD holders and experienced professionals. Senior Fellows are people with deep experience, such as a PhD, an industry track record or strong publications.
What do fellows do? They carry out a research project, alone or in a team, under the guidance of a mentor. Past work falls into three areas: technical AI safety, technical AI governance, and policy and governance. Mentors come from groups such as the University of Warwick, the Institute for Progress, the Centre for Long-Term Resilience and Meridian Labs. Each fellow gets a weekly one-to-one with a mentor and weekly support from a research manager. There are also workshops, speaker sessions and a cohort of about 40 peers. The goal is a research output, such as a paper or policy brief, plus clear next steps for your career.
The package goes beyond cash. Pivotal covers compute for research and weekday lunch and dinner at LISA. Fellows are not employees. They sign a stipend agreement instead.
The fellowship has a track record. About 160 people from eight cohorts have gone on to groups such as the UK AI Security Institute, GovAI, METR, Apollo Research, Google DeepMind and the Institute for AI Policy and Strategy. Some alumni have founded their own organisations, and many have started PhDs. The programme used to run for nine weeks. It now runs for 15.
For African applicants, two practical points matter. Pivotal says meeting UK entry rules is your own job, though it shares an information document and can sometimes help with the process. Remote participation is rare and mostly for people with a strong research record. If you cannot be in London for the full period, say so in your form.
The application has several stages. First comes a written form that takes about an hour or two, including your mentor choices. Shortlisted applicants then do a short video interview, a mentor-specific task of one to three hours and a personal interview of 30 to 45 minutes. Pivotal’s advice is to pick mentors with care. A strong match with a mentor often counts for more than your overall background.
Why it matters: AI tools are reaching African classrooms, banks, farms and governments. The people who study how to keep that technology safe are mostly based elsewhere. A paid fellowship that welcomes many disciplines is a rare way for African researchers, policy staff and engineers to join that work and build links with leading groups.
What to watch next: the 1 November deadline, the mentor list for the Q1 2027 cohort, and the research outputs fellows publish after April.
Sources: Pivotal Research call for applications, shared via LinkedIn; Pivotal Research, AI Safety Fellowship; EA Forum post on the Q1 2027 fellowship.
A note for startup founders: Don’t practice on investors. Practice on AI first. RaiseReady by Index Prima is an AI fundraising simulator for African founders: upload your pitch deck, get a readiness score and face a tough AI investor before the real meeting. Find this post, and every post we publish, at raiseready.indexprima.com.
