Readings & resources
Resources to help you get started in AI safety and understand the shifting landscape.
All
A curated collection of courses, readings, tools, and communities for understanding AI safety and finding ways to contribute.
Courses
Structured learning paths for building a clear foundation in AI safety, strategy, and effective action.
LAISH Fellowship Coursework
Based on the AI Safety Atlas, this is the core coursework used in our Foundations of AI Safety Fellowship.
BlueDot Impact Courses
Cohort-based courses in AI safety, governance, and strategy for people exploring high-impact work.
Lens Academy Courses
A growing catalogue of courses on AI safety, alignment, control, and related technical topics.
Short Course on AGI Safety
An introductory technical course from DeepMind researchers on the foundations and open problems of AGI safety.
Effective Altruism Introductory Program
A guided introduction to using evidence and careful reasoning to find effective ways of improving the world.
Europe
Analysis and practical guides focused on Europe's role in governing increasingly capable AI systems.
Europe 2031
A scenario examining the geopolitical and economic choices Europe may face as AI capabilities accelerate.
The Race Worth Winning
An argument for how middle powers can shape a safer and more prosperous age of machine intelligence.
EU Artificial Intelligence Act
An independent resource hub for understanding the EU AI Act, its timeline, obligations, and implementation.
Technical
Accessible starting points for understanding neural networks, modern AI systems, and technical safety debates.
Neural Networks
A visual video series explaining how neural networks learn, from first principles through modern language models.
Attention Is All You Need
The landmark research paper that introduced the Transformer architecture underlying today's leading language models.
Don't Worry About the Vase
Frequent, detailed commentary on AI capabilities, safety research, policy, and industry developments.
Governance
Scenarios, strategies, and political analysis for navigating the transition to advanced AI.
Strategic Visions in AI Governance
A report mapping strategic approaches to governing the development and deployment of advanced AI.
AI 2027
A detailed scenario exploring how rapid progress toward superhuman AI could unfold and reshape global affairs.
AI 2040: Plan A
A proposal for international transparency and cooperation on a safe path toward highly capable AI.
ChinaTalk
A newsletter and podcast on technology, national security, and Chinese policy, with extensive coverage of AI competition.
Communities & Opportunities
Places to find training, roles, collaborators, and thoughtful discussion across the AI safety community.
80,000 Hours Job Board
A curated job board for roles tackling pressing global problems, including technical and governance work in AI safety.
AI Safety Training Directory
A directory of courses, fellowships, research programs, and workshops for developing AI safety skills.
Atlas of AI Safety Training
A searchable database of AI safety training programs and learning opportunities around the world.
LessWrong
An online community for improving reasoning and discussing AI alignment, forecasting, and long-term challenges.
Effective Altruism Forum
A forum for research, debate, and coordination among people working on high-impact global problems.
Share a reading or resource
Memos, briefs, and projects from the community are welcome.