Scientists using LLMs will ‘do more, less well’, modelling study predicts

www.nature.com · 2026-08-01 · by Long Nguyen ·generic
Research suggests that the combination of incentives to publish and the use of large language models will lead to more papers, but they will be less refined.

Scientists using LLMs will ‘do more, less well’, modelling study predicts

A person holds up a smartphone displaying an AI chatbot app against a colourful, futuristic-style blurred background

The rising use of artificial-intelligence tools by scientists could increase their productivity, but this might come at the cost of quality.Credit: Getty

Scientists who use large language models (LLMs) to help them with their research will spend less time refining their work and instead jump quickly to fresh projects, according to a modelling study1.

Adoption of LLMs will cause scientists to “do more, less well — rather than the same amount, better”, write the authors. But they say that such artificial-intelligence tools do not bear all of the blame for that outcome.

The results reflect a flawed incentive system in science that prioritizes quantity over quality, says study co-author Carl Bergstrom, a biologist at the University of Washington in Seattle. Scientists are incentivized to churn out papers, he says, and LLMs help them to reach this ever-growing quota. “LLMs are rarely the problem themselves,” says Bergstrom. “LLMs hold up a mirror to problems that we already have.” The study was posted on the arXiv preprint repository on 19 July and has not yet been peer reviewed.

Foraging for papers

To predict how LLMs will change scientific productivity, the authors broke the research process into discrete phases. First comes a discovery phase when scientists brainstorm hypotheses and conduct initial experiments to determine the value of a project. Next, there is a two-part development phase that consists of required work (such as creating figures and drafting papers) and discretionary development (such as conducting follow-up experiments and polishing writing).

The authors drew on methods from optimal-foraging theory — which analyses how an animal maximizes energy gain while conserving resources — to analyse how scientists will reallocate their efforts after incorporating LLMs into their work. The authors’ model assumed LLMs are functioning at their best — cheap, fast and accurate.

AI linked to explosion of low-quality biomedical research papers

The modelling predicts that LLMs can speed up all phases of the scientific process, but that this acceleration won’t result in better papers. Faster discovery and required-work phases mean that researchers can churn out papers more quickly than they could without the help of LLMs, and the pressure to publish means there is little incentive to spend extra time polishing those papers.

Quality not quantity

Enjoying our latest content?

Log in or create an account to continue

School of Information Science and Technology invites highly qualified candidates to fill multiple tenure-track/tenured faculty positions. Shanghai (CN) ShanghaiTech University

www.nature.com · 23 images · 124 links

Comments (0)

No comments yet