OpenAI Targets a Fully Automated AI Researcher by 2028

OpenAI named its "north star": a fully autonomous AI researcher capable of independently tackling complex problems in math, physics, biology, and policy. Chief scientist Jakub Pachocki outlined a two-stage roadmap - an AI research intern by September that can handle tasks taking a human a few days, scaling to a full multi-agent research system by 2028. The interim milestone is telling: they're not aiming for breakthroughs first, they're aiming for reliable delegation. GPT-4, reasoning models, and Codex are described as the building blocks for systems that work extended stretches with minimal human oversight. Google DeepMind's AlphaProof and Anthropic's internal research agents have been moving in the same direction, but OpenAI is the first to publicly frame autonomous research as the company's primary objective rather than a side project.