tech
March 20, 2026
OpenAI is throwing everything into building a fully automated researcher
An exclusive conversation with OpenAI’s chief scientist, Jakub Pachocki, about his firm's new grand challenge and the future of AI.
TL;DR
- OpenAI is prioritizing the development of an autonomous AI researcher as its "North Star" for the coming years.
- The company plans to release an AI research intern by September and a fully automated multi-agent research system by 2028.
- This AI researcher will be capable of tackling problems too large or complex for humans, spanning math, physics, life sciences, business, and policy.
- Key to this goal are advancements in reasoning models, agents, and interpretability, building on technologies like GPT-4 and Codex.
- While acknowledging risks such as AI going off-rails or being misused, OpenAI is focusing on monitoring techniques like chain-of-thought.
- The development of such powerful AI raises significant societal and governmental challenges regarding control and deployment.
- OpenAI's chief scientist, Jakub Pachocki, believes this technology will be transformative even if AI does not match human intelligence in all aspects by 2028.
Continue reading the original article