TLDR
New writing tools have made more papers and reviews appear, but the writing and variety of ideas have worsened. Unless universities and journals change how they reward work, the system may favor producing more work instead of producing better work.
Summary
This content was automatically synthesized by Credo's AI models directly from the original source text.
AI summaries can make mistakes — double-check important details against the original source.
1 Study Aim
The authors examine how artificial intelligence (AI), computer systems that generate or analyze content, is changing journal submissions and peer review. They ask whether AI, combined with publish-or-perish incentives (pressure to produce many publications), is increasing output while reducing quality. The study avoids deciding which uses of AI are appropriate because adoption is still new. The research asks whether academic publishing is becoming busier without becoming better.
2 Study Design
The study analyzes 6,957 initial submissions and 10,389 textual reviews at Organization Science from 2021 through 2026. The authors compare writing quality in submissions and reviews associated with AI-generated writing against human-generated writing. They also examine topical diversity, meaning the range of subjects discussed in reviews. Conversations with editors across scientific fields provide broader context, although the authors do not claim these observations represent every journal. The researchers compared thousands of papers and reviews before and after AI tools became widely available.
3 Findings
The study reports a 42% increase in submission volume after ChatGPT’s release in late 2022, alongside declining writing quality. The authors estimate that AI-generated writing explains nearly all of these changes. AI-generated reviews also became more common, showed lower writing quality, and covered a narrower range of topics than human-generated reviews. Editor conversations suggest similar pressures across disciplines. The authors argue that current tools and publish-or-perish incentives are pushing research toward more output rather than better work. They recommend adapting institutions and reward systems if AI is to support innovation. More papers and reviews are being produced, but the overall system may be losing quality and variety.