OpenAI Announces Progress on Automated Research Intern System

OpenAI has successfully developed an 'Automated Research Intern' system capable of independently executing multi-day research tasks under human guidance. As of mid-August, the research department's AI agents were handling a workload 3.1 times greater than that of human researchers. This development accelerates an 'AI training AI' feedback loop, where AI-generated code and experimental results lead to faster research progress and the creation of more capable models.