Over the past century, the focus of scientific practices has shifted from purely intellectual exploration to problem-solving, leading to uneven development in scientific knowledge. Our analysis of 41 million research articles over the past six decades reveals this trend of uneven development, with atypical papers representing complementary innovation becoming the majority and displacement papers representing substitutive innovation decreasing to the minority. While AI can enhance human memory capacity, it may not necessarily accelerate progress in canonical concepts without changing the agenda of science and its organization.

翻译：过去一个世纪以来，科学实践的重心已从纯粹的知识探索转向问题解决，导致科学知识呈现不均衡发展。我们对过去六十年间4100万篇研究论文的分析揭示了这一不均衡发展趋势：代表互补型创新的非典型论文已成为多数，而代表替代型创新的位移论文则减少为少数。尽管人工智能能够增强人类记忆容量，但若未能改变科学议程及其组织形式，其未必能加速经典概念的进步进程。