Newsletter

Join the FAR.AI Newsletter

Please check your spam/junk folder if you don't see our confirmation email in your inbox.

Past newsletter

2026 Q1

The Deception Problem: New Results and Open Questions

Alignment Workshop coming to Seoul; sign up to workshops on AI control & verification; our new research on AI deception & more; and, we're hiring!

2024

AI Safety as a Global Public Good

International Dialogues, Alignment Workshops, Jailbreak-Tuning, AI Agents Planning, and More!

2025 Q4

From Discovery to Deployment: Shaping Safer AI Systems

We're hiring! Also, our upcoming events, research on persuasion, honesty, and sandbagging, and a tender from the European Commission!

2025 Q3

Scaling Our Impact, Accelerating Critical Research

We found critical vulnerabilities in GPT-5 and Opus 4’s safeguards against misuse, including for chemical, biological, radiological, and...

2025 Q2

Building Bridges: From Research to Global Workshops

Red-teaming frontier models, using lie detectors to make AI more honest, connecting technical experts and policymakers, and bringing our Alignment Workshop to Asia. Plus, we’re hiring!

2025 Q1

AI Safety: From Research to Global Action

Paris AI Security Forum, London Control Workshop, Jailbreak-Tuning Demo, Expanding Our Team, and More!

Get research updates

Subscribe for summaries of our research in adversarial robustness, interpretability, and more!

Join the community

FAR.Labs in Berkeley hosts events like weekly AI safety seminars featuring top researchers, and invites technologists to collaborate in our research hub.