Should Researchers Write Papers for AI Instead of People?

A provocative proposal from 37 researchers at leading universities and technology companies is challenging the fundamental purpose of scientific publishing. In a new paper, they argue that scientists should shift their focus away from writing for human readers and instead create content that is optimized for AI agents. The authors contend that as AI systems become increasingly involved in research processes, the literature itself must evolve to be machine-readable and actionable. The core of their argument rests on efficiency. If AI agents are to assist in literature reviews, hypothesis generation, and experimental design, they need structured data that can be parsed and processed without ambiguity. Human-oriented papers, with their narrative flow and implicit context, are often difficult for machines to interpret accurately. By designing papers with AI in mind, the research community could accelerate discovery and reduce the time spent on manual data extraction. However, this proposal has ignited a heated debate. Critics worry that prioritizing machine readability could alienate human researchers, particularly those in fields where narrative explanation is crucial for understanding complex concepts. There are also concerns about accessibility, as not all researchers have the tools or skills to produce AI-optimized formats. The discussion raises fundamental questions about the purpose of scientific communication: is it for advancing knowledge, or for facilitating automated processing? As AI continues to permeate every aspect of research, this debate will likely intensify. The challenge lies in finding a balance that serves both human understanding and machine efficiency, ensuring that scientific progress remains inclusive and transparent.