The AGI Safety and Alignment team at Google DeepMind is Hiring (July 2026)

Alignment Forum · Seb Farquhar · 2026-07-31

Google DeepMind's AGI Safety and Alignment Team announces multiple open research and engineering roles, distinguishing its approach from Anthropic and OpenAI safety teams by emphasizing knowledge production over advocacy, a focus on future high-capability systems, and a hiring process that requires candidates to use agentic coding assistants.

Open original ↗

Appears in

AI Alignment Research Attracts Major Funding While Challenging Core Assumptions

Extraction

Topics: agi-safetyai-hiringgoogle-deepmindalignment-research

Claims

ASAT is actively hiring for multiple research and engineering roles across all areas of its safety portfolio.
ASAT's approach differs from other frontier lab safety teams by preferring knowledge production over advocacy and focusing on systems more capable than current models.
All ASAT engineering interviews allow unlimited agentic coding assistance, and candidates who do not use coding agents are expected to fail.
ASAT expects its AI control measures to become 'load-bearing for safety in the near future,' signaling a shift in organizational priority.
Google's institutional reputation means ASAT's safety work has greater policy impact than equivalent work at other labs.

Key quotes

All of our engineering interviews allow coding agents (and we expect that most people who choose not to use coding agents will fail them).

We prefer knowledge over advocacy. Everyone wants AI systems to be safe. Often, the challenges are factual disagreements... and execution.

We expect our control measures in particular to become load-bearing for safety in the near future.