Researcher, Recursive Self-Improvement Safety

The role

This person will perform hypothesis-driven research and develop interventions or control systems to mitigate risks from accelerated AI development. The role involves designing pre-deployment risk assessments, control measures, and scalable oversight for model misbehavior.

What you will do

• Consider future problems OpenAI might face and how to prepare for them.
• Turn open-ended objectives into concrete research directions.
• Execute quickly by building and iteratively improving scrappy prototypes.
• Secure buy-in from other staff and communicate work clearly.
• Collaborate with or manage other staff as needed.
• Design experiments and evaluations to understand model misalignment.
• Prototype technical mechanisms for verifying compliance with AI safety agreements.
• Track progress toward automation of technical staff.
• Maintain and strengthen RSI safety cases.

Required qualifications

• Exceptional technical executor.
• Strong strategic and research taste.
• Passionate about mitigating the risks associated with recursive self-improvement.
• Driven by a desire to do work that positively impacts the future of AI development.

Nice to have

• Work in ML research, AI alignment, or AI verification

About the company

OpenAI is an AI research and deployment company dedicated to ensuring that general-purpose artificial intelligence benefits all of humanity. They seek to safely deploy AI through their products.