OpenAI Plans to Achieve Autonomous AI Researcher by 2028
OpenAI has set an internal goal to develop a truly autonomous "legal AI researcher" system by March 2028.
This system will be capable of independently undertaking large research projects, including autonomous experiments, data analysis, and scientific discoveries, with the aim of achieving AI intern-level capabilities by September 2026.
In terms of market mechanisms, AI developers and research institutions are accelerating the allocation of computing resources and capital to the OpenAI ecosystem due to the potential for autonomous research. Driven by events, funding is shifting from labor-intensive R&D to AI agents and multi-agent systems, benefiting OpenAI and its computing partners, while traditional academic and corporate research teams face productivity pressures.
Source: Public Information
ABAB AI Insight
Sam Altman has previously publicly accelerated the roadmap for 2025, promoting OpenAI's transition from the GPT series to o1/o3 reasoning model iterations and leading collaborations with giants like Microsoft for computing power. The 2028 goal continues his long-term strategy of shifting AI from a tool to an autonomous scientific discoverer.
In terms of capital pathways, OpenAI is mobilizing millions of GPU resources to build autonomous research systems through massive GPU cluster investments and multi-agent architecture development, motivated by the goal of achieving a self-accelerating cycle (AI helping to develop better AI), while establishing a technological leadership position for future AGI regulation and commercialization, attracting more sovereign funds and corporate strategic investments.
Similar to DeepMind's early AlphaFold breakthroughs in protein folding, and the evolution of multi-agent systems from simple tasks to complex projects in 2025, OpenAI is currently in a critical expansion phase transitioning from human-led AI R&D to AI autonomous scientific research.
Essentially, this represents a technological substitution: autonomous AI researchers will replace some functions of human researchers in the scientific discovery process, driven by the rapid iteration of reasoning models and agent system capabilities, forcing R&D resources to shift from human recruitment to computing power and data infrastructure, achieving a structural reconstruction from labor-intensive research to AI-driven exponential discoveries.