AI Reverses 60-Year Decline in Solo Scientific Authorship

A new study titled 'Return of the solo author' by researcher Akira Matsui has uncovered a surprising shift in scientific publishing. By analyzing over 300 million works across 26 fields, Matsui found that the decades-long decline in solo-authored papers—a trend that had been steadily moving toward larger teams—halted and partially reversed following the public release of ChatGPT in late 2022. This reversal is not uniform: it is strongest in fields where a coauthor's work can be readily replaced by AI, such as computational and digital disciplines, and weakest or absent in fields that rely on physical collaboration, like experimental lab work.

The study delves deeper into who is driving this change. The uptick in solo authorship is not coming from new entrants or shifts in field composition. Instead, it appears among established authors who previously only coauthored papers, as well as long-standing researchers and newcomers alike. Their solo works stay close to their own coauthored research but are narrower in scope and lean heavily into computational topics. Because a solo paper lacks credited human coauthors, this offers direct empirical evidence that generative AI can substitute for scientific labor. Matsui argues that this signals a reconfiguration of cognitive labor within papers rather than merely a change in team size—suggesting AI is reshaping the very division of labor in science.