Resources

The latest edition of Stanford HAI’s independent annual assessment of the state of artificial intelligence.

The 2025 edition of the AI Index — benchmarks and analysis across research, performance, economy, and policy.

Stanford HAI’s independent annual assessment of the state of artificial intelligence — the program home.

A benchmark tracking how the largest banks are adopting and deploying artificial intelligence.

An aggregated index of artificial-intelligence metrics and reference material.

Independent benchmarks and comparisons of AI models and the providers that serve them.

The artificial-intelligence section of arXiv — the open-access preprint server that feeds our Library’s research stream.

An AI-powered academic search engine — the source of the citation, venue, and abstract data on our paper pages.

A catalog of model metadata, licensing, and benchmark scores — a metadata layer behind our Models and Benchmarks.

A global index of artificial-intelligence readiness, capability, and adoption.

A comprehensive, living database of risks from artificial intelligence, maintained by MIT.

A public, searchable record of harms and near-misses caused by AI systems in the real world.

Anthropic’s framework for managing catastrophic risks as AI capabilities scale.

OpenAI’s framework for tracking and preparing for frontier model risks ahead of deployment.

Perplexity’s research blog — engineering and applied-AI work from the search company.

Stanford HAI’s plain-language glossary of core artificial-intelligence terms and concepts.

OpenAI’s hub for its safety practices, research, and policy commitments.

An industry body advancing the safe development of frontier AI through shared research and best practices.

An association working on standards, governance, and responsible practice across the AI industry.

A nonprofit focused on AI safety, ethics, and the responsible stewardship of advanced systems.