Overview
- OpenAI published a Preparedness-team job listing this month seeking a researcher to prepare for recursive self-improvement and offering $295,000 to $445,000 in base pay.
- The posting names concrete technical priorities: defending models from data poisoning, building tools to interpret model reasoning, and measuring how AI automates technical work inside the company.
- The Preparedness team already covers automated red-teaming, biological and chemical risk assessments, and threats from agentic AI, and the new role extends that portfolio toward self-improving systems.
- The hiring follows public timeline and capability signals from industry leaders and studies that show rapid model gains and explicit company goals to automate research work over the next few years.
- If firms succeed at automating research, companies will need new detection and containment tools and the role could shape how OpenAI tracks internal automation and protects jobs, safety, and external stakeholders.