A Summary of Recent Work (July 2026)

By Rohin Shah and Seb Farquhar.

It’s been nearly two years since our last major update in August 2024 and we wanted to share another recap of our recent work. Things have changed a lot since then. We are now fully in the midgame, and focus more on landing things in production. We are the AGI Safety and Alignment Team (ASAT), the main group at Google DeepMind working directly on technical approaches to existential risk from AI systems. Last year we published An Approach to Technical AGI Safety and Security, which remains the best place to read our overarching vision. Norms around chain of thought. Our impression is that…

saved by

• Yixiong Hao
• Eliana Du

related reading

• A Pragmatic Vision for Interpretability — AI Alignment Forumalignmentforum.org
• SFT Drives Gemini’s Safety Properties — LessWronglesswrong.com
• AI in 2025: gestalt — LessWronglesswrong.com
• Recommendations for Technical AI Safety Research Directionsalignment.anthropic.com
• [2507.11473] Chain of Thought Monitorability: A New and Fragile Opportunity for AI Safetyarxiv.org
• Core views on AI safety: When, why, what, and how \ Anthropicanthropic.com
• I lead AGI safety at Google DeepMind – here's the view from the inside | Rohin Shahyoutube.com
• Core views on AI safety: When, why, what, and how \ Anthropicanthropic.com
• A Pragmatic Vision for Interpretability — LessWronglesswrong.com
• AI Control: Improving Safety Despite Intentional Subversion — LessWronglesswrong.com
• Security incident disclosure — July 2026huggingface.co
• AI Control: Improving Safety Despite Intentional Subversion — AI Alignment Forumalignmentforum.org