Resources
Courses, readings, and fellowships in AI safety and governance.
Get started
At Cornell, you can start with CS 1998 and our reading groups and workshops. You can learn about AI safety with other students or study independently using our public course materials.
BlueDot Impact runs facilitated online courses in technical AI safety, frontier AI governance, and biosecurity. You can study the core ideas through readings, exercises, and discussions with a cohort.
ARENA develops practical research-engineering skills in deep learning, interpretability, reinforcement learning, and evaluations. You can work through the notebooks independently or apply to a four-to-five-week cohort in London. You’ll need a background in Python and the relevant mathematics. Explore the self-study notebooks.
CAMBRIA is CBAI’s three-week AI safety bootcamp using ARENA materials, focused on interpretability and reinforcement learning. It has offered cohorts in Cambridge, Massachusetts, and New York City. You can check CBAI for the next cohort and location.
Iliad Intensive is a four-week, full-time course on the mathematical foundations of AI alignment, including deep learning theory, agent foundations, and interpretability. It is aimed at people with strong mathematical backgrounds, with cohorts in London and Berkeley.
Fellowships & training
You can explore opportunities in research, policy, community building, and starting new organizations. You’ll find eligibility, locations, and application dates on each program’s website.
Technical safety & research
- MATS. Mentored research in alignment, transparency, and security.
- Anthropic Fellows. A funded four-month program for empirical AI safety research with Anthropic mentorship.
- CBAI AI Safety Research Fellowship. Funded ten-week research in Cambridge, Massachusetts, across technical safety and governance. Undergraduates are encouraged to apply.
- CHAI Research Internships. Paid, mentored research in human-compatible AI at UC Berkeley. Open to students and experienced technical applicants.
- LASR Labs. Team-based technical AI safety research in London.
- Astra Fellowship: Empirical Research. Funded, mentored research at Constellation on alignment, control, evaluations, and scalable oversight. Astra also offers a strategy and governance stream.
- Iliad Fellowship. Three months of full-time, mentored mathematical AI alignment research for PhD-level researchers or equivalent experience.
Policy, governance & law
- GovAI Fellowships. Research and applied tracks in AI governance and policy.
- GovAI Research Scholar. A paid twelve-month role for developing research or applied projects in AI governance.
- IAPS Fellowship. AI policy and strategy projects with expert mentorship.
- LawAI Fellowships. Research at the intersection of law and AI governance.
- Horizon Fellowship. Policy training and funded placements in Washington, DC, focused on AI and other emerging technologies.
- RAND CAST Fellows. Technology, security, and policy research at RAND.
- Talos Fellowship. A pathway into European AI policy careers.
- Vista AI Law and Policy Fellowship. Mentored AI law and policy research for students and recent graduates.
- FAI Fellowships. Policy training, including programs for conservative policy professionals.
- TechCongress. Places early- and mid-career technologists in US congressional offices to work on technology policy.
- Google Public Policy Fellowship. Student placements at public-interest technology policy organizations. AI-related projects depend on the host.
- Aspen Science and Technology Policy Fellowship. Policy training and an optional mentored project for US-based science and technology experts.
- Tech Policy Press Fellowship. Editorial mentorship for reporting and analysis on technology and democracy, including AI governance.
Generalists, founders & community building
- Generator Residency. Kairos and Constellation’s funded three-month residency in Berkeley for building AI safety programs, infrastructure, and organizations. Includes mentorship and support entering generalist roles.
- MATS Founding and Field-Building. A twelve-week track for founders, generalists, and organizers to launch AI safety initiatives with experienced mentors. Projects can become new organizations or help existing ones grow.
- Astra Fellowship: Strategy and Governance. Funded work on AI strategy and governance, with mentorship and incubation support for new projects or organizations. Suited to independent thinkers with strong AI safety context.
- Pathfinder Fellowship. Kairos’s mentorship and activity funding for students running university AI safety or policy groups. Open to organizers worldwide, including those starting a new group.
Cross-disciplinary & flexible pathways
- SPAR. Part-time, remote projects in technical safety, policy, and biosecurity.
- Pivotal Research Fellowship. Mentored work across AI safety and governance.
- ERA Fellowship. Research across technical safety, governance, and technical governance.
- PIBBSS / Principles of Intelligence. Interdisciplinary approaches to intelligence and alignment.
- CAIS AI and Society Fellowship. Funded research in San Francisco on AI’s economic, legal, geopolitical, and social impacts. For advanced researchers.
- UChicago XLab Fellowship. Summer research on AI safety and nuclear security for students and early-career researchers. See the latest cohort details.
- Vitalik Buterin Fellowships. Support for doctoral and postdoctoral research in AI safety and governance.
Library
Browse papers, articles, videos, and curricula by topic.