Follow the people shaping AI safety.

Research leadGoogle DeepMind

Mechanistic interpretability, practical research walkthroughs, and ways for new researchers to get oriented.

Research scientistGoogle DeepMind

Alignment research, agent security, and clear synthesis of new technical work across the field.

Alignment researcherAnthropic

Frontier-lab alignment research, scalable oversight, and automated methods for making safety research move faster.

Executive directorAlignment Research Center

Scalable oversight, reinforcement learning from human feedback, alignment strategy, and risk estimates.

Founder and CEOMETR

Independent frontier-model evaluations and measurements of whether models can complete longer, autonomous tasks.

Co-founder and CEOApollo Research

Scheming, deception, model monitoring, and independent evaluations of frontier AI systems.

Co-founder and CEOFAR.AI

Robustness, red teaming, external scrutiny, and moving promising safety research into wider use.

Executive directorCenter for AI Safety

Safety benchmarks, catastrophic risk, robustness, and research intended to make advanced systems more reliable.

Chief scientistRedwood Research

AI control, capability elicitation, empirical threat models, and forecasts about advanced AI behavior.

CEORedwood Research

AI control, security research, technical strategy, and candid discussion of research bets and field needs.

AI policy researcherCSET

AI governance, external oversight, transparency, and the national-security dimensions of advanced AI.

FounderAVERI

AI policy, verification, auditing, responsible deployment, and the institutions needed around frontier systems.

Co-founder and head of policyAnthropic

AI policy, compute, research trends, and a broad weekly view of developments through Import AI.

ResearcherMETR

AI timelines, risk models, frontier-risk reporting, and careful thinking about research prioritization.

Associate professorUC Berkeley

Machine-learning robustness, evaluations, forecasting, and technically grounded analysis of frontier systems.

Independent writerDon’t Worry About the Vase

Fast-moving synthesis of AI policy, lab news, safety arguments, and debates happening across the field.