OpenAI Sets Sight on Fully Automated AI Researcher by 2028
OpenAI has adopted the development of a fully autonomous agent-based research system as its primary objective for the coming years, starting with an AI research intern.
OpenAI has unveiled a new core pursuit: developing a fully automated, agent-based AI researcher designed to solve large, intricate problems independently.
According to details shared by OpenAI Chief Scientist Jakub Pachocki, the San Francisco-based company considers this project its primary "north star" over the next few years.
The Road to Automated Discovery
Rather than jumping directly to an end-to-end autonomous scientist, OpenAI's roadmap outlines an incremental progression. The immediate milestone is an "autonomous AI research intern." This intermediate system will be tasked with solving a focused set of specific research problems.
The AI intern will serve as a foundational testbed and precursor for a significantly more capable, fully automated multi-agent architecture. OpenAI expects this advanced multi-agent system to make its debut in 2028.
Scaling Frontier Ambitions
The initiative reflects the industry's broader shift from simple conversational chatbots toward proactive, goal-oriented AI agents capable of carrying out sustained, multi-step tasks without continuous human guidance.
If realized, an autonomous research system could dramatically accelerate scientific exploration, engineering breakthroughs, and AI model development itself, marking a significant milestone in agentic artificial intelligence.
