Directory
- Center for Human-Compatible AI (CHAI)
UC Berkeley research center under Stuart Russell, working on AI systems that stay under human correction.
Research · 148 people · 1 open job
- Mox
San Francisco coworking and events space for AI safety and adjacent communities.
Field building · 141 people
- Horizon Institute for Public Service
Places emerging-technology experts into US federal agencies, Congress, and think tanks through a fellowship.
Governance · 106 people · 2 open jobs
- Leverhulme Centre for the Future of Intelligence (CFI)
Cambridge interdisciplinary center on the long-term future of artificial intelligence.
Research · 82 people
- Centre for the Governance of AI (GovAI)
Oxford-based research organization on AI governance, and one of the field's main sources of policy research.
Governance · 74 people · 4 open jobs
- MATS
Twelve-week fellowship pairing scholars with alignment researchers as mentors, running in Berkeley and London. One of the field's main pipelines.
Training · 74 people · 2 open jobs
- Principles of Intelligence (Princint)
Nonprofit bringing researchers from the natural and social sciences into AI safety. Runs the PIBBSS fellowship, coordinated research programs, and an in-house research team.
Training · 73 people · 1 open job
- Center for Security and Emerging Technology (CSET)
Georgetown policy research center producing data-driven analysis of AI, national security, and export controls.
Governance · 71 people · 3 open jobs
- Centre for the Study of Existential Risk (CSER)
Cambridge research center studying risks that could end or permanently curtail humanity's prospects.
Research · 67 people
- Schmidt Sciences
Philanthropy funding science across five centers, one of them AI and Advanced Computing, whose grantmaking includes safety-focused calls and the AI2050 fellowship.
Funding · 66 people · 1 open job
- Forecasting Research Institute (FRI)
Research institute on forecasting methods applied to existential risk, co-founded by Philip Tetlock.
Research · 60 people
- Oxford Martin AI Governance Initiative (AIGI)
Oxford Martin School research program on AI governance and international coordination.
Governance · 57 people
- Future of Life Institute (FLI)
Nonprofit on catastrophic and existential risk, running grant programs, fellowships, and public campaigns.
Funding · 52 people · 11 open jobs
- FAR.AI
Research nonprofit on adversarial robustness, evaluation, and alignment. Also runs the FAR.Labs coworking space.
Research · 51 people · 16 open jobs
- Alliance to Feed the Earth in Disasters
Research nonprofit on feeding the world through global catastrophes that cut off normal agriculture.
Research · 50 people
- Partnership on AI
Multi-stakeholder nonprofit developing shared practice on responsible AI across companies and civil society.
Governance · 47 people
- Rethink Priorities
Research and implementation organization working across animal welfare, global health and development, survey and data analysis, and worldview investigations. Its Special Projects team fiscally hosts other organizations.
Research · 46 people
- Americans for Responsible Innovation (ARI)
US advocacy nonprofit working on emerging-technology policy in the public interest, including frontier AI.
Governance · 45 people · 2 open jobs
- Ada Lovelace Institute
Independent UK institute researching how data and AI systems affect people, and what regulating them should look like.
Governance · 44 people
- Apart Research
Distributed research organization running hackathons and mentored studies that bring new people into technical AI safety.
Research · 43 people · 3 open jobs
- Apollo Research
London research and auditing organization evaluating whether frontier models scheme or deceive their evaluators.
Research · 42 people · 12 open jobs
- METR
Evaluates whether frontier models have dangerous autonomous capabilities, and publishes the methodology.
Research · 40 people · 9 open jobs
- Epoch AI
Research group measuring and forecasting AI progress: compute, data, algorithms, and their trends.
Research · 37 people · 7 open jobs
- Mila - Québec AI Institute
Montreal deep learning institute founded by Yoshua Bengio, with a growing AI safety research group.
Research · 35 people · 1 open job
- Constellation
Berkeley research center, published as Constellation Institute, offering desks, fellowships, and residencies to people working on AI safety.
Field building · 34 people · 2 open jobs
- Institute for AI Policy and Strategy (IAPS)
Washington think tank on frontier AI policy, covering compute governance, security, and international strategy.
Governance · 32 people
- AI Objectives Institute
Research nonprofit applying mechanism design and cooperative AI to the question of what AI systems are pointed at.
Research · 31 people
- ERA:AI Fellowship
Fully funded Cambridge summer fellowship on frontier AI risk and its governance.
Training · 30 people
- Saxe Lab
UCL Gatsby lab under Andrew Saxe, published as the UCL Theory of Learning Lab, building mathematical theory for how learning works in neural networks and in brains.
Research · 30 people
- AE Studio
Development studio with an in-house alignment team working on neglected technical and policy approaches.
Research · 28 people · 7 open jobs
- The Future Society
Nonprofit working on AI governance in the EU, the US, and international forums.
Governance · 28 people · 1 open job
- Concordia AI
Beijing-based advisory on safe AI development, and a main English-language source on Chinese AI safety work.
Governance · 28 people
- Machine Intelligence Research Institute (MIRI)
Long-running nonprofit on superintelligence risk, now focused on policy and communicating the case publicly.
Research · 25 people · 2 open jobs
- Forethought
Research foundation on how to steer the long-run trajectory of advanced AI, co-founded by Will MacAskill.
Research · 25 people
- Longview Philanthropy
Advises major donors on giving to reduce risks from advanced AI, nuclear weapons, and biological threats, and runs its own grantmaking funds including the Frontier AI Fund.
Funding · 25 people
- MIT Algorithmic Alignment Group
MIT CSAIL group under Dylan Hadfield-Menell on human-aligned AI, with published research areas spanning value learning, incentives, recommendation systems, debugging, policy, and human-AI interaction.
Research · 25 people
- Softmax
Research organization developing an account of organic alignment, co-founded by Emmett Shear.
Research · 23 people
- Transluce
Nonprofit lab building open tools for understanding and steering model behavior at scale.
Research · 22 people · 4 open jobs
- ControlAI
Advocacy organization campaigning for binding regulation of frontier and superintelligent AI.
Governance · 22 people · 1 open job
- Collective Intelligence Project
Research organization designing democratic mechanisms for deciding how AI systems are governed.
Governance · 21 people · 1 open job
- UChicago Existential Risk Laboratory (XLab)
University of Chicago lab on existential and global catastrophic risk across AI, biosecurity, nuclear weapons, and climate. Runs a summer research fellowship and student working groups.
Research · 21 people
- Center on Long-Term Risk (CLR)
Research nonprofit on cooperation and conflict between advanced AI systems, aimed at worst-case outcomes.
Research · 20 people · 1 open job
- Arcadia Impact
UK nonprofit running AI safety programs, covering model evaluations and student field-building.
Field building · 20 people
- Bau Lab
Northeastern University interpretability lab under David Bau, studying the internal structure of deep networks.
Research · 20 people
- Timaeus
Research nonprofit that applied singular learning theory and developmental interpretability to alignment. Its own site states it has merged into Resolution, where its open roles now sit.
Research · 20 people
- TruthfulAI
Empirical research group under Owain Evans on truthfulness, situational awareness, and emergent misalignment.
Research · 20 people
- Anthropic
Frontier AI lab whose research agenda includes interpretability, scalable oversight, and a published responsible scaling policy.
Research · 19 people · 652 open jobs
- SaferAI
French organization building risk management frameworks for AI companies, and rating companies against them.
Governance · 19 people · 3 open jobs
- Foresight Institute
Runs grant programs and workshops across AI safety, neurotech, and other frontier technologies.
Funding · 19 people · 1 open job
- LawZero
Nonprofit lab founded by Yoshua Bengio developing safe-by-design AI as an alternative to agentic systems.
Research · 18 people · 14 open jobs