Chinese mathematician presents 1,000 unsolved problems to test artificial intelligence

A prominent Chinese mathematician has published a collection of more than 1,000 unsolved mathematical problems to test both human scholars and modern software models. As reported by the South China Morning Post, Nanjing University professor Sun Zhiwei created these challenges to evaluate whether automated systems can truly innovate or if they merely excel at processing existing knowledge.

For non-experts, this initiative underscores a major boundary in modern technology: moving beyond routine pattern matching toward genuine intellectual discovery. Advanced computer systems can easily solve textbook problems, but advanced mathematics requires forming original hypotheses. To ensure fair testing, researchers cut off the software's web access so the models cannot simply look up existing research or give up when encountering an unsolved problem.

It is still unclear whether current software models can independently resolve complex, open conjectures without human assistance. Furthermore, if a computer does generate a novel solution, human experts will still have to carefully audit the proof to confirm its mathematical validity. This raises a compelling question: if algorithms eventually master original mathematical reasoning, how will that redefine the way we educate students in logic and problem solving?