Resources

Clear, accessible video explainers on AI alignment and why it's a hard problem — no technical background required.

A Latter-day Saint apostle on faith, moral agency, and artificial intelligence.

MIRI's concise case for why building smarter-than-human AI with today's understanding risks human extinction.

A detailed, researcher-authored scenario forecasting how transformative AI could unfold over the next few years.

A policy proposal for an international agreement to responsibly delay superintelligent AI until 2040.

An overview of why AI safety may be one of the world's most pressing problems, and how a career addressing it could look.

Structured, cohort-based online courses for going deeper on AI safety, alignment, and governance.

The hub for technical AI alignment research and discussion.

Community blog on rationality and AI safety — much of the field's writing lives here.

Regular updates on AI safety research, policy, and news.

Technical alignment research and interpretability work from Redwood Research.

Newsletter on responsible AI integration, with balanced perspectives on hype, safety, and social impact.

Reporting and analysis on the power and politics of transformative AI.

A weekly rundown of AI research and policy from Anthropic co-founder Jack Clark.

A dense, closely-read weekly roundup of AI developments and commentary.

Career-focused updates on AI safety and other pressing global problems.

In-depth interviews on the world's most pressing problems, including AI safety, and how to build a career addressing them.

Long-form interviews with AI researchers and other thinkers on where the technology is headed.

Conversations with AI builders and researchers on the frontier of the field and what it means for the world.

An auto-updating board of AI safety jobs, fellowships, grants, events, and courses.

A part-time, remote research program pairing aspiring researchers with AI safety and governance mentors.

A fully-funded, 9-week London fellowship for research in AI safety, AI governance, and biosecurity.

A fully-funded Cambridge fellowship for researchers and entrepreneurs working to understand and reduce risks from frontier AI.

Places fully-funded fellows with emerging-tech expertise in US federal agencies, congressional offices, and think tanks.

A 13-week, London-based technical AI safety research programme, working in small mentored teams.

A fully-funded Cambridge/Boston research fellowship covering interpretability, governance, and other AI safety topics.

A funded fellowship placing journalists in newsrooms to strengthen reporting on AI.

A paid, four-month program for researchers and engineers to work on AI safety projects alongside Anthropic's research teams.

A funded program for external researchers to pursue rigorous research on the safety and alignment of advanced AI systems.

A four-week, in-person course on theoretical approaches to AI alignment, for those with a strong math background.

A curated job board for roles in AI safety, spanning technical, policy, and operations work.

A job board of high-impact roles, including many in AI safety, vetted for their potential to do good.