Nicholas Polson and the AI-Generated Academic Paper Surge

AI-Driven Volume in Academic Publishing

Nicholas Polson has authored 258 academic papers in 2026 so far, a volume of output that has sparked significant debate regarding the integrity of academic research and the impact of generative AI on scholarly communication. This surge in production highlights a growing tension between the ability to generate content at scale and the traditional standards of peer review and academic validation.

The Distinction Between Preprints and Peer-Reviewed Journals

A critical point of contention in the discussion is whether these works are formally published in peer-reviewed journals or hosted on preprint servers.

IIUC there are not 258 academic papers published in serious peer review journals, only 258 "preprints" in a free preprint server. It's like posting in WordPress but using a PDF instead of html.

Preprint servers allow authors to upload manuscripts before they undergo formal peer review. This distinction is vital because preprints do not carry the same weight as peer-reviewed research, as they have not been vetted by independent experts in the field. The ability to upload PDFs to these servers allows for a rapid increase in the volume of work available, but not necessarily the increase in quality or intellectual contribution.

Co-authorship and Validation

Despite the high volume, some co-authors have defended the practice. Vadim Sokolov, a co-author on some of the papers, has stated that they reviewed at least one of the generated papers, asserting that they are not "fake papers."

However, other observers have questioned the quality of the writing and the utility of such a high volume of output. Some have noted that generative AI models, while capable of producing convincing formats, often struggle with the rigor required for high-level academic work, such as that produced by PhD or postdoctoral students.

Motivations and Theoretical Implications

Observers have proposed several theories regarding the motivation behind this massive output:

• Testing Limits: Some suggest this may be a "Everest effect," where the author is simply testing the limits of what is possible with new technology.
• Idea Squatting: Others argue it may be a form of "squatting," where publishing a manuscript quickly establishes a claim to an idea, even if the work is not fully developed.
• Psychological Impact: There are suggestions that the use of LLMs can create a feedback loop of "AI-psychosis," where the dopamine hit of seeing speculative ideas expanded into a convincing format encourages further production without critical review.

The Future of Academic Review

The rise of AI-generated preprints has led to calls for a better system of academic review. The current system relies heavily on unpaid reviewers who are already overwhelmed by low-quality submissions. The prospect of AI-generated submissions being met with AI-generated reviews could further erode the value of the publishing metric as a measure of academic quality.

If AI submissions, then why not go a step further and have AI reviewers? Perhaps this will help bring more scrutiny to the metric of publishing quantity not being as valuable as it has been treated.

This trend suggests a shift where written media may lose value, and the focus may return to in-person communication and public speaking as primary means of verifying intellectual contribution.