Why location matters
Being physically near other safety researchers dramatically accelerates your work. Hallway conversations, whiteboard sessions, and spontaneous collaborations are how many of the best ideas in alignment have started. If you can, consider visiting one of these hubs.
Bay Area 4
Toronto 1
New York 1
Blackpool 1
Cambridge 1
London 2
Zurich 1
Cape Town 1
Singapore 1
Sydney 1
Select a pin to jump to a city. For a fuller directory, including local meetup groups, see AISafety.com’s community map.
San Francisco Bay Area
4 spacesLondon, UK
2 spacesCambridge, UK
1 spaceZurich, Switzerland
1 spaceSingapore
1 spaceToronto, Canada
1 spaceNot near a hub? Start online.
You don't need to relocate to start contributing. The two best on-ramps are a structured course and a weekend research sprint β both are remote, open to newcomers, and give you something concrete to show for your time.
BlueDot Impact β AI Safety Fundamentals
The standard entry course. Free, remote, and cohort-based: you work through a structured curriculum in a small facilitated group over several weeks. Most people in the field have either taken it or recommend it β and the cohort itself becomes your first network.
Apart Research hackathons
Weekend research sprints on a set alignment problem, run remotely with teams from all over the world. A rare chance to do real research with essentially no barrier to entry β and the strongest submissions sometimes turn into papers.
Beyond those, the AI Alignment Forum and the EleutherAI Discord are good places to lurk and then participate. When you're ready, visiting a hub for even a few weeks can be transformative.