Will the widespread use of large language models in scientific writing undermine scientists’ critical thinking?

2026/06/03 by Lucas M. Bietti, Adrian Bangerter · 1 voice
Arts and Humanities · Decision Sciences · Social Sciences · #Academic Writing and Publishing #scientometrics and bibliometrics research #Climate Change Communication and Perception

paper · doi:10.1371/journal.pbio.3003801

Abstract

Artificial intelligence (AI) is rapidly transforming scientific writing by expanding access and efficiency, yet it risks decoupling writing from thinking. Scientific writing is a core cognitive and epistemic practice that must be cultivated and preserved alongside AI use.

Citations

• Delving into LLM-assisted writing in biomedical publications through excess vocabulary
• Quantifying large language model usage in scientific papers
• Technical Report on the Pangram AI-Generated Text Classifier

Discussions

• Generative AI can make it quicker and easier to write scientific research papers. But as Lucas Bietti & Adrian Bangerter argue in this Perspective, unchecked #AI use risks decoupling #writing from #th [bsky, 37 points, 3 comments]

Related

• Scientific Writing, Generative Artificial Intelligence, and the Non–Native English Speaker
• Advancing the Scientific Method with Large Language Models: From Hypothesis to Discovery
• Scientific Reliance on AI may Reflect Broader Failures in Science Communication
• Generative AI Robs Students of the Joy of Learning
• The artificial intelligence revolution...in unethical publishing: Will AI worsen our dysfunctional publishing system?