Two sites track nearly every AI safety opening, so check them first. The 80,000 Hours job board lists roles, fellowships, internships, and PhD positions, and AISafety.com's events and training page lists programs currently open for applications. Everything below is our shorter, opinionated list. If you're weighing options, 80,000 Hours offers free one-on-one career advising for people working on AI safety and policy.

For readings and lists of open research problems, see the AI Safety Reading List.

<aside> 🧭

Plan first

Consider filling out this career memo before you apply anywhere. It walks you through what you want from the next few years, and your mentor is a good person to ask for feedback on it.

</aside>

<aside> 📖

Courses

• BlueDot Impact runs free online AI safety courses (AGI Strategy, Technical AI Safety, Frontier AI Governance, and Biosecurity) as one-week intensives or five-to-six-week part-time cohorts, and we think they're the best introduction to the field. Put spar-fall-2026-mentee in the “Where did you hear about this course?” question when you apply. The mentee guide FAQ has the details.
• ARENA teaches the ML engineering behind technical AI safety, self-paced or in person in London.
• AI Safety, Ethics, and Society is a free textbook covering the field's core arguments. </aside>

<aside> 🎓

Fellowships and research programs

Technical AI safety:

• MATS
• Anthropic Fellows Program
• Constellation Visiting Fellows and the Astra Fellowship
• ERA Fellowship
• CHAI

AI policy and governance:

• GovAI fellowships
• RAND Technology and Security Policy Fellows
• Talos Fellowship
• Paragon Policy Fellowship
• Horizon Fellowship
• Vista Institute for AI Policy programs
• Tarbell Fellowship

A longer list lives in this Airtable of fellowship opportunities. Many programs run on fixed cycles, so note the deadlines now even if you'd apply after the round.

</aside>

<aside> 🏢

Where people work

The job board above is the complete list. A few organizations we'd point you to if you want somewhere to start looking:

• Technical: METR, Redwood Research, Apollo Research, FAR.AI, Alignment Research Center, Center for AI Safety, Anthropic
• Policy: GovAI, UK AI Security Institute, CAISI, CNAS, European AI Office, Coefficient Giving
• Academia: many AI safety PhDs happen outside dedicated labs, and AI safety master's programs are multiplying. Two labs with strong track records are the Algorithmic Alignment Group at MIT and the Bau Lab at Northeastern. </aside>

<aside> 💸

Funding for independent work

Plenty of people work on AI safety as independent researchers or on small projects, funded by grants rather than a salary:

• Long-Term Future Fund
• Survival and Flourishing Fund
• AI Safety Fund
• Coefficient Giving requests for proposals
• BlueDot Rapid Grants </aside>

<aside> 🎪

Conferences and events

• EA Global and EAGx, where much of the field meets each year
• NeurIPS and the other major ML conferences, if your project is technical (SPAR can fund travel to present accepted SPAR research, see the mentee guide FAQ)
• Global Challenges Project workshops, which Kairos runs for students
• AISafety.com's events and training page for everything else </aside>