一项针对4130万篇学术论文的分析显示，使用AI工具的科学家在发表数量、引用量和晋升速度上显著优于同行，但导致科学探索范围收窄和原创性下降 [1]。该研究由芝加哥大学社会学家 James Evans 领导，结果于1月14日发表在《自然》期刊上 [1]。分析数据集包含1980年至2025年间发表的英文论文 [1]。数据显示，使用AI的科学家平均发表论文数量是不使用者的3倍 [1]，获得引用量接近5倍 [1]。此外，使用AI的科学家比同行早1到2年成为团队领导者 [1]。研究指出，这种个人职业发展与集体科学进步之间的张力，源于当前学术评价体系对速度和规模的奖励机制 [1]。
A study led by sociologist James Evans from the University of Chicago reveals a tension between individual career advancement and collective scientific progress in the age of artificial intelligence. The research, published on January 14 in Nature, analyzed over 41 million English-language academic papers released between 1980 and 2025 [1].
Scientists who utilized AI tools demonstrated significant advantages in their professional metrics compared to those who did not; they published three times as many articles and received nearly five times the number of citations on average [1]. Furthermore, researchers employing these technologies became team leaders one to two years earlier than their peers [1]. However, this efficiency came at a cost: the use of AI tools resulted in a narrowed scope of scientific exploration and a decline in originality within the field [1].
Evans attributes this divergence between personal success and broader discovery to current academic evaluation systems that prioritize speed and volume over depth or novelty [1].