● Covered by 1 source · 1 reportMedium impact
Study shows 32% of new arXiv papers may be AI-generated
A study analyzed 12,750 arXiv papers and found that approximately 32% read as machine-written, post-ChatGPT. The research used a calibrated detector to establish a baseline and highlight the rising trend of AI-generated content in academic papers.
Key points
- 32% of new arXiv papers flagged as AI-generated.
- Study analyzed 12,750 papers from various fields.
- Computer science shows the highest AI-content at 65%.
Study Methodology
The study assessed the full text of 12,750 papers on arXiv from 2021 to 2026, focusing on detecting AI-generated writing.
A detector with a 0.4% false-positive rate was used, calibrated based on papers prior to the introduction of ChatGPT.
Key Findings
Results indicate a steady false-positive rate of 0.4% in 2021 and 2022, followed by an increase to 32% of papers flagged as AI-written after ChatGPT's release.
The share peaked near 39% in early 2026, varying significantly across academic fields.
Field-Specific Trends
Computer science papers had the highest AI-written flag rate at approximately 65%.
Mathematics showed the lowest rate, around 0.7%, which the authors noted may be due to specific contextual factors inherent to the field.
Implications of Findings
This study provides a clearer understanding of AI's impact on academic writing quality and quantity, raising questions about authenticity in published research.
The detection methodology and results serve to inform discussions on the role of AI in academia and future research credibility.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
The daily brief
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Today's brief
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's brief
New every morning, and the back catalogue is archived by date.
Reporting from
A study analyzed 12,750 arXiv papers and found that approximately 32% read as machine-written, post-ChatGPT. The research used a calibrated detector to establish a baseline and highlight the rising trend of AI-generated content in academic papers.