OpenAI Chief Research Officer Mark Chen said last week that artificial intelligence could produce scientific discoveries worthy of a Nobel Prize within the next two years. Speaking at the company's Seoul office during the International Conference on Machine Learning, Chen pointed to recent AI-driven breakthroughs in chemistry and mathematics as evidence that the pace of discovery is accelerating.
"Within the next two years, artificial intelligence will produce groundbreaking innovations worthy of major awards," Chen said, as reported by South Korea's Chosun Daily. "Whether it will win a Nobel Prize is uncertain, but it will make discoveries of immense value to the world." Chen, who was born in 1990 to a Taiwanese family in the U.S., earned degrees in mathematics and computer engineering from MIT and has volunteered as a coach for the U.S. team at the International Mathematical Olympiad for more than a decade.
AI's recent track record in science
In 2024, researchers behind Google DeepMind's AlphaFold received the Nobel Prize in Chemistry for using AI to predict protein structures. More recently, a Chinese AI system solved a decade-old mathematics problem in 80 hours with minimal human involvement, according to a team at Peking University. In May, an OpenAI chatbot solved the unit distance problem-a puzzle first posed by Hungarian mathematician Paul ErdΕs in 1946-producing a proof that several mathematicians said would likely qualify for publication in a leading journal if a human had written it. British mathematician Timothy Gowers, a Fields Medal winner, said: "No previous AI-generated proof has come close" to meeting that standard.
These advances are part of a wider shift in AI for Science & Research, where machine learning models are increasingly used to accelerate hypothesis generation and experimental design.
The next horizon: AI researching AI
"At this rate, AI will soon become intelligent enough to research AI itself," Chen said. He added that AI-driven discoveries, particularly targeted treatments for diseases, could deliver enormous benefits to humanity.
Math education remains essential
When asked whether children should continue studying mathematics and science in the AI era, Chen emphasized the value of foundational skills. "Mathematics teaches how to think. It shows how to reason, solve problems, and cultivate resilience. At a fundamental level, math must remain a core subject," he said.
Why this matters for Science and Research
For researchers, AI is becoming a lab partner that can substantially shorten the time from hypothesis to result. The ability to solve decades-old problems in days signals that AI tools will soon be indispensable for both theoretical and applied work. Scientists should watch for new AI-assisted discovery platforms and consider how to integrate these tools into their workflows while preserving the critical thinking skills that Chen and others say remain essential to the field.
Your membership also unlocks:
