Anthropic Residency (paid AI safety research placements)
Anthropic · San Francisco (remote possible for shorter placements); UK posting
Home › AI safety
34 open · 22 fellowships & residencies · 9 grants · 3 competitions & prizesFellowships, grants, research sprints and red-teaming challenges for people working to make advanced AI go well, from technical alignment and interpretability to policy, law and security.
Some links on this page are affiliate links. We may earn a small commission at no extra cost to you.
Not sure you qualify? Try Am I eligible?Five quick questions and we'll pick the open opportunities that fit you best.Anthropic · San Francisco (remote possible for shorter placements); UK posting
Centre for the Governance of AI (GovAI) · Washington DC, USA
Centre for the Governance of AI (GovAI) · London, UK
Tarbell Center for AI Journalism · Newsroom placements (see organiser page)
Kairos · Online
AI Safety Camp · Online
Oxford Martin AI Governance Initiative, University of Oxford · Oxford, UK
Apart Research (with Security Level 5) · Online
Talos Network · Brussels / Europe
International Programme on AI Evaluation · Online (+ Valencia, Spain)
AI Security Bootcamp · London, UK
Apart Research · Online
Iliad · London / SF Bay Area
SPAR · Remote
Oxford Martin AI Governance Initiative, University of Oxford · Oxford, UK
Institute for Law & AI (LawAI) · USA
Institute for Law & AI (LawAI) · Cambridge, UK / Washington DC
Coefficient Giving · Global
Constellation · Berkeley, CA, USA
Forethought · Oxford, UK or Berkeley, US
Global Technology Risk (GTR) Foundation · Global
London AI Safety Research (LASR) Labs · London, UK
MATS Research · Berkeley, CA, USA / London, UK
PIBBSS · See organiser page
Anthropic · UK / USA / Canada
BlueDot Impact · Global
BlueDot Impact · Global
Gray Swan AI · Online
Iliad · London / SF Bay Area
Lightcone Infrastructure · Global
Mercor · US or UK
Coefficient Giving · Global
Survival and Flourishing Fund (SFF) · Global
EA Funds · Global
Plain-English guides to the kinds of funding on this page, including recent articles.
Several AI safety and alignment programmes still have open applications this winter, but a few of the tightest close in the next week. This is a snapshot of what is still open as of 18 November 2026, who each one is for, and which recently advertised rounds have already closed.
By advance.to · 4 min readA growing number of programmes pay people to spend a few weeks or months on AI safety, governance or policy research. They are not interchangeable. Some pay a senior salary and want established leaders, some pay nothing and take almost anyone, and some are policy placements. Here is how to tell them apart.
By advance.to · 3 min readA growing number of funders give grants, not fellowships, to people and new organisations working on AI safety and catastrophic-risk reduction. Most of the programmes we track are rolling, so you can apply when you are ready. This article explains how they differ.
By advance.to · 4 min readPolicy and law fellowships pay you to spend a summer, or a few months, on AI governance, not to train a model. They are easy to confuse with AI safety research residencies. This article separates the main policy programmes that are open now, and what actually differs between them.
By advance.to · 4 min readRed-teaming challenges pay you to find ways an AI system fails: jailbreaks, prompt injection, weak attributions, bad threat models. Some run all year; others are a single weekend. The rules and the prize pool are the first things to read, not the last.
By advance.to · 3 min readMost people who move into AI safety research do it through a structured programme: a mentor, a project and a few months of funded time. The programmes look similar from the outside, but they differ a lot in pay, time commitment, location and how selection works. This article compares the main ones using their own pages, checked on 18 November 2026.
By advance.to · 5 min read
Your First AI Funding Application shows you which AI funding fits you and how to write an application that gets read. Then, every Monday, one short email with the deadlines closing in the next 14 days.