Mathematical discoveries in the age of artificial intelligence (amu.edu.pl)

Dr Bartosz Naskręcki and Prof. Ken Ono published a commentary in Nature Physics, "Mathematical discovery in the age of artificial intelligence," exploring how AI is reshaping mathematical research. The piece surveys two complementary trends: the automation of proof work (via proof assistants and automated theorem provers) and the rise of AI as a creative partner that helps generate conjectures, spot patterns, and suggest new directions. The authors argue this shift is more than a productivity boost—it's changing how mathematics is done, evaluated and taught, with implications for what counts as a verified result. Technically, the commentary highlights opportunities and risks: machine learning–guided search and symbolic–numeric hybrids can accelerate discovery, but they also raise verification, reproducibility and interpretability challenges that traditional mathematical standards must address. The authors call on the academic mathematics community to lead the development of responsible practices—standards for formal verification, curated datasets, reproducible pipelines and governance around transparency and credit. For the AI/ML community, the piece emphasizes fertile ground for tooling that integrates formal methods with statistical models and for collaborative workflows that preserve mathematical rigor while leveraging AI’s pattern-finding strengths.