Jeremy Horpedahl tells the story:
Nicholas Polson has, by my count using his SSRN page, already written 258 working papers in 2026 alone. He’s already written (or at least published to SSRN), six papers today, August 26, 2026.
OK, but who am I to talk?—I’ve written over 200 blog posts this year. But wait:
These aren’t just short notes. Most of the papers are of normal academic length: 32 pages, 27 pages, 58 pages. . . . Obviously the research productivity of Polson and his co-author Sokolov is aided by AI. . . . I have seen any academic, at least not in economics, that has really pushed it to the limit.
Horpedahl writes:
Read any single paper, and it feels like just a normal academic paper, the kind of thing that an academic might work on for a few months.
I wanted to see if I shared that judgment so I clicked through to the list of Polson’s papers on SSRN and looked for something interesting . . . ok, here’s something. It’s called Theories of Human Connection, and . . . ulp! It’s 80 pages long. The paper’s subtitle is “An Interdisciplinary Synthesis Across Economics, Psychology, Biology, Philosophy, Game Theory, and Spiritual Tradition.”
But let’s take a look. The abstract on SSRN starts like this:
Human connection is the most studied and least integrated phenomenon in the social sciences. Every discipline that examines intimate relationships captures something real that the others miss, yet no existing framework holds all dimensions in simultaneous view. This book synthesises fourteen thinkers across biology, psychology, economics, game theory, communication theory, existential philosophy, and spiritual tradition into a unified account. Morris established that the need for physical touch is an evolutionary drive as fundamental as hunger, with a biologically ordered sequence of escalating vulnerability whose disruption produces intensity without depth. Bowlby showed how early caregiving creates invisible templates governing adult intimacy, encoded in the nervous system before language exists. Becker revealed that partnerships generate value neither partner could produce alone, but assumed partners are interchangeable — an assumption Frankl demolishes by showing that the meaning generated by shared life cannot be transferred, and that meaning multiplies rather than adds to life satisfaction, explaining widespread disconnection in the wealthiest societies in history. Von Neumann’s game theory explains how mutual self-protective withdrawal, individually rational for each partner, produces the disconnection neither intended. Bateson identifies the communication structure that accelerates this collapse: contradictory demands that make any response wrong. Gottman’s laboratory models predict separation with over ninety percent accuracy from the ratio of positive to negative interactions. The Hindu philosophical tradition adds the final dimension: partnership as a laboratory in which selfishness and fear are progressively revealed and surrendered.
We argue that most relationship failures are not failures in one dimension but misidentifications of which dimension is actually in play, and that effective intervention requires the multi-dimensional map this synthesis provides.
From the preface:
The thinkers assembled here — Becker, Bateson, Von Neumann, Schelling, Keynes, Morris, Vaughan, Frankl, Maslow, Gottman, Yogananda, Vivekananda, Maharaj, Polson, Thomas, and Paltrow — did not, for the most part, know each other’s work. They worked in different centuries, different countries, different intellectual traditions. What unites them is that each identified a dimension of human connection that the others left in shadow.
Gottman, huh? The name rings a bell . . . He’s the guy who conned Malcolm Gladwell and various media outlets—maybe he conned himself too—into believing that he could predict divorces with 94% accuracy.
I searched for Gottman in the document and found a whole chapter on that bullshit! You can click through for yourself, it’s chapter 10. Here’s a key bit:
I wonder what prompts were used by Polson and his coauthor to write this article. I guess the prompts did not include, “Evaluate implausible claims skeptically.”
They bring it all together in Chapter 13, “The Cumulative Model: A Unified Theory of Human Connection”:
But let’s not forget “The Master Equation” on page 75:
Jesus Christ. I’ve heard that Cambridge University has an open position in their school of education . . . this kind of thing would fit in very well there, no?
In all seriousness, no, I don’t think this “feels like just a normal academic paper, the kind of thing that an academic might work on for a few months.” At least, not the sort of thing a non-bullshitting academic might write.
I know Nick Polson—he’s a statistician, and he’s done lots of solid work over the years! What happened here?
Here are a few possibilities:
1. It’s an experiment or a joke. But if it were a joke I’d think there’d be some internal clues, no? It’s hard to imagine playing the whole thing straight. And if it were an experiment, I’d expect they’d all read like straight-up statistics papers, nothing so obviously bogus as “Theories of Human Connection.”
2. Someone else is impersonating Polson. Seems unlikely, but it’s possible. It might not even be personal. Maybe all this is an experiment, not by Nick but by someone else who programmed a chatbot to choose the name of a successful academic and then spew the internet with papers attributed to him. If so, how horrible.
3. “Intellectual squatting.” That’s the conjecture of Michael Makowsky, who writes:
The nice version is it’s putting out a series of half-baked papers in the hopes of establishing a property right to the underlying ideas at an earlier stage of the research process than previously possible. The less generous interpretation is it’s dumping a series of haystacks on the plains and laying claim to the needles probabilistically within each.
I guess . . . but what does Nick ultimately get out of it? Invitations to speak at more conferences?? I don’t get it.
Makowsky writes:
Imagine you are a person who has highly esoteric, potentially important ideas every day. Many of those ideas you suspect, based on some combination of experience and ego, are new in at least one dimension. You would like to get credit for that newness. For being first. What’s the problem?
The problem is that scholarship remains more perspiration than inspiration. Having a new idea is great, but it takes years to work through the nuance in sufficient detail that you can convince your peers of the coherence and originality of the contribution. During the minutes each day you are not working on this singular project you have the inspiration for other ideas, sometimes multiple within a single day. How frustrating is the proposition that someone else gets credit for the originality of contribution just because they had time to reveal it to the world while you were embroiled in your investigation of what is only one of your many score ideas!?
Ah, but meta-level inspiration has struck you! What if you took each one of those ideas, spent an hour curating a series of prompts around it, and then let Chat GPT (or another LLM) fabricate an entire research paper around it? . . .
But . . . that’s what blogging’s all about! Often when I have an idea or a reaction, I blog it. No need to pipe it through a chatbot; I’ll just save the cycles and post it right here.
To return to the 256 working papers, I can think of one more motivation:
4. Education. This seems like the most plausible explanation to me. Polson has had an active research and teaching career, and he’d like to share his insights with a broader audience than the readers of his published papers and the students at the University of Chicago business school. And one way to reach people is . . . econ preprints! So Nick picks 258 interesting topics, writes some prompts for each, and produces the articles. I guess he’s programmed a bot to do this. He just feeds it the prompts and the bot writes the paper and posts it directly to SSRN.
That could explain the mystery of how that ridiculous 80-page article with “The Cumulative Model: A Unified Theory of Human Connection” (shades of Stephen Wolfram!) ended up there. Not only can’t you expect an author to write 258 articles of that length in less than a year, you can’t expect him to read all of them too. The content of that bizarre article could be as much a surprise to Polson as it was to me.
This then raises a question: setting aside the motivations of Polson (or his impersonator), do these 258 papers have any value?
It’s hard for me to answer this question, given that I’ve only looked at one of them. My guess is that the net value of the papers is negative, in that the amount of time that people (including me) have wasted going through them outweighs any positive contributions that might have been there.
My suggestion
Here’s what Nick could do on this, which could have value: Take these 258 prompts and write an article (himself, not using the chatbot) explaining why he thinks these ideas are important. Aki and I wrote a paper a few years ago, What are the most important statistical ideas of the past 50 years?. Nick could write something similar: What are the 258 most important things in statistics to learn today? Or something like that. I’m not saying it would be easy—it would take more effort than programming a chatbot to spam SSRN—but valuable products often take work to produce. Nick has tenure and could set aside the time to do it.
Also I’d recommend withdrawing all those papers from SSRN. Withdrawing 258 papers seems like a lot of work, but I’m sure he could easily program a bot to do the job.
P.S. There’s a further twist: there are two accounts for Nicholas or Nick Polson at the University of Chicago business school; see this comment thread. This would seem to be consistent with the “social experiment” hypothesis (if Nick decided to set up a separate account to play around with) or the “impersonation” hypothesis (if the bot that wrote and posted these papers was not created by Nick at all). The whole thing remains a mystery to me.
P.P.S. OK, I did a bit more nosing around.
SSRN allows you to list the papers in time order. If you go to Nick’s SSRN page linked from his website, you’ll see 16 papers, with the first (“The Impact of Jumps in Volatility and Returns”) being posted on 1 Jan 2001, then others through the next two decades, with the most recent being “Deep Learning in Characteristics-Sorted Factor Models,” posted on 23 Sep 2018 and last revised 26 Jun 2023.
If you go to the SSRN page with all the fake papers, it starts with “Kramnik vs Nakamura or Bayes vs p-value,” posted 7 Dec 2023. It’s a badly written paper—I’m guessing not AI, just text by a non-English-speaking author that was not ever checked by native speakers before posting. This rings a bell . . . I actually have a blog post on this paper, scheduled to appear next year. Next on the list is a 25-page paper, “AI and Vivekananda,” posted 5 Mar 2024, then a gap of two years until another AI-related paper appeared on 9 Mar 2026, then on 11 Mar 2026 came the aforementioned “Theories of Human Connection.”
So, yes, Polson has two SSRN pages, but they have no overlap in time. He also has papers on Arxiv, including the intriguingly-titled “Bayes with No Shame: Admissibility Geometries of Predictive Inference,” dated 24 Aug 2026 . . . Hey, that’s just 3 days ago! Oddly enough, I can’t find this one on SSRN.
But what about the article itself? I don’t have the patience to read it, but I did catch that it mentions the martingale property, which is one of my current interests—that’s cool. But, just flipping through, it looks much more substantive—much more like a real scientific paper—than that horrible “Theories of Human Connection” thing. This could be a tribute to the power of modern chatbots to create something so convincing.
P.P.P.S. Update here: The incredible shrinking SSRN page.
P.P.P.P.S. Vadim Sokolov, coauthor of several of the chatbot-assisted papers, comments here. Assuming this comment itself is legitimate, those 258 papers were not an experiment or a hoax, nor was there any impersonator, nor was it intellectual squatting. Rather, Polson and his coauthors just had 258 different things to say, and they felt the best way to do so was using a chatbot to write tens of thousands of words on these topics.
Also, according to Sokolov, they did not program a bot to do this. He reports that at least one of the papers went through multiple rounds of review, so even if a chatbot was involved in that one, the human contribution was much more than inserting a prompt. He says that some of the papers “are much newer and were developed with much more extensive AI assistance. Modern AI made it possible to develop and complete this material at a speed that would previously have been impossible.”
The fallacy in this statement by Sokolov is the idea that using a chatbot to turn a few hundred words of prompts into ten thousand words of text is a way to “complete this material.” Nothing’s being “completed” here in any intellectual sense; the process just adds pages and pages of froth. I think that Josh’s analogy of this to the Sorcerer’s Apprentice is apt.
P.P.P.P.P.S. Polson appears to think he’s proved the Riemann hypothesis (see comments here and here). I would think this would cause some concern among his collaborators.
A post about 258 papers with negative utility has positive value for me: if this guy submits any papers to my journal I’ll know not to spend any time on them. In all seriousness: crap like this can only clog up the system. What should follow is all penalty, no reward.
I love AI. The early impact of AI is on many processes and institutions which were losing their utility anyways. AI and LLMs just reveal how pointless publishing has become (other than a currency for academic credentials)
I would love AI to be the straw that breaks the camels back for the academic publishing enterprise. I think such AI related papers will increase to the point where it poisons the well. The average quality of a publication was down anyways and AI might be just what we need to get rid of this conventional model.
First I heard complaints about authors using AI. Then reviewers and editors. Ultimately where’s the human in the loop? Presumably in the future people will get their answers by asking AI instead of reading papers anyways.
I agree and I think your comment applies quite widely. Teaching has been thrown into turmoil – how to evaluate student work and insure it was theirs? So much of what we did was broken that AI is exposing all the faults. AI is not the problem, it is the canary in the mine. This doesn’t mean that AI isn’t dangerous, but I think most of the things people focus on are problems overdue to be addressed. To take another: AI threatens to undermine democratic processes – so much misinformation, false videos, messages designed to take advantage of AI ‘capabilities.’ These are real problems, but the democratic process (in the US, at least) was already broken. So much reliance on donations, so much special interest money, and our 2 party oligopoly make a mockery of representative democracy. AI makes all of this worse, but it was already bad.
If you know him, then why not ask him?
Jd:
I don’t know him well. It’s been several years since I’ve talked with him or interacted with him in any way, and it seemed awkward to ask him. I guess I could’ve just shot him a quick email with a “Whassup with that?”
Less awkward than writing a public blog post about him?
“In all seriousness, no, I don’t think this “feels like just a normal academic paper, the kind of thing that an academic might work on for a few months.” At least, not the sort of thing a non-bullshitting academic might write.”
I don’t know Andrew. I’ve seen some pretty atrocious academic work written by seemingly sincere academics that is of equivalent quality to what I’ve sampled of Polson’s AI generated work. I’d hate to pick on the humanities and the softer social sciences (anthropology, some psych stuff, sociology)… god knows econ, poli sci, and even some things from the harder sciences are not great… but there are entire academic fields where what Andrew quoted would be perfectly adequate or maybe even top notch output.
I go with number 1. His own website doesn’t raise any red flags and he has a long history of published research. Without speaking to its quality, the impossible volume of manuscripts in the last year looks like an experiment to see what AI does to academic research. I wouldn’t call it a joke since I think the impacts are serious business. But I can’t see how any human can average more than one (fairly detailed) paper per day. Even with my poor sleep habits it seems physically impossible without using AI. And I can’t imagine that he is thinking that his reputation or career will be enhanced by inundating his resume with this stuff. I vote for experiment.
Dale:
I agree that “experiment” is a likely possibility: once the bot is set up to write papers and post them automatically, why not just set it to submit hundreds of them at once? It’s not clear to me what’s being tested in such an experiment, but maybe that’s the point. I guess we’ll find out soon.
I will say that, if it is an experiment, props to Nick for putting himself out there, in the same way that authors of autofiction risk embarrassing themselves in order to convey a larger truth.
I’m wondering if this is a Sokol-esque play to see how far one can get with crap research in an effort to expose it. I can’t tell if this is actually a likely possibility or if I’m just coping because I quite like Polson’s previous work.
Think the sorcerers apprentice.
Josh:
Based on the comment by Sokolov in this thread, I’m thinking that your Sorcerer’s Apprentice theory might be right. Polson and his collaborators got a hold of this powerful tool and couldn’t let go.
re: red flags
He posted a 10 page proof of the Riemann Hypothesis way back in 2017 to math.GM (https://arxiv.org/abs/1708.02653) with a long revision history (and a retraction? in the Brazilian Journal of Probability and Statistics) that goes up to July of this year. Seems suspicious, even if this is his (very strange) idea of a fun prank or something. IMO it feels like a undercurrent of crankery has gotten amplified by LLMs even though he’s clearly a serious and competent researcher in (non-RH) areas.
The literature was already like this, you should already be assuming its 99.99% junk.
People have to figure out its independent replication and surprising predictions that matter. Using lots of big words and esoteric symbols is not a good proxy. In fact accessible jargon and shorter documents are better heuristics.
The astronomers say there is a lunar eclipse tonight, and Im confident I can look up in the sky to see it myself without even knowing how to read. *That* is the indication of good science.
This is more depressing than most AI misuse stories. And the 13-dimensional mathematical “model” is hilariously stupid, regardless of whether it was devised by man or machine. I share your puzzlement about why anyone would churn out this flood of AI-generated articles. Of your proposed explanations:
1. (A joke / experiment) I don’t think so, for the reasons you gave. But it’s plausible. I hope this is the reason!
2. (impersonation) No; I’m sure he would have posted a statement by now.
3. (squatting) — plausible, but see below.
4. (education) I don’t see how Polson would think that these papers have educational value, or even if they did, that simply pointing students to them would aid their learning. (It’s not like there’s a shortage of papers out there.)
Two other possibilities:
5. Delusion. Given that AI tools are brilliant at so many things, Polson has come to believe that they are brilliant at everything, and all these papers describe true great insights into economics (or whatever field idiodic scientism about 13-dimensional behavior models fits into). He thinks his prompts have led to over 200 breakthroughs.
6, or 3b. I think the “squatting” hypothesis makes the most sense, but not with the aim of getting conference invitations, but of filling the future AI training data with himself. As Tyler Cowen sometimes annoyingly writes, he’s writing for the AIs. Scraping the SSRN pages, the models will find a prolific Nicholas Polson, who can’t be ignored simply by reason of volume if some hapless student is seeking to learn about 21st century business (?) thinking. Polson will be immortal, like the deadbeat fathers of 20 children who, almost by definition, are more biologically fit than those of us with just a few.
“the [chatbot] models will find a prolific Nicholas Polson, who can’t be ignored simply by reason of volume”
That has to be it! He’s getting in on the proverbial ground floor. Since chatbots also feed on their own tail, his importance could grow exponentially over time.
Which begs the question: at what point do chatbot citations get considered in tenure review? Users posed prompts on things they care about that resulted in your papers being cited 6,000 times last year, that’s gotta mean something, right? And heck, why stop there? Let’s rate university departments by how many chatbot citations their professors garner.
People tell me I need to flag my sarcasm. The previous paragraph was sarcasm! Unfortunately, it is painfully close to the various Red Queen dilemmas that have already formed around AI.
One problem with this strategy: all those papers are in public domain.
No; copyright is retained:
License Information
The copyright holder has granted SSRN a license. All rights reserved. No reuse allowed without permission.
There is no copyright holder, assuming the papers were indeed AI generated.
To be clear for those who aren’t familiar. Its far from established about AI, but precedent says that there must be a substantial input from an actual human in order for a work to be copyrightable. If the work is basically AI generated then it isnt copyrightable and so the supposed copyright licensed to SSRI simply doesnt exist.
It’ll take some kind of court cases to fully establish this but I think its safe to say that the precedent is strongly against the idea of a chatbot generated work being copyrightable unless someone sat there prompting for individual paragraphs and then rewriting them and soforth… where theres substantial human input.
I can’t tell if you’re being entirely serious (you only noted that the second-to-last paragraph was sarcastic), but as Andrew noted, Polson was already a relatively well-known and respected statistician – he was well beyond the “ground floor” by this point. I have no idea why he would do this.
I’ve read a fair bit of his work myself over the years. This is a super disappointing thing to read.
If it’s #2 (someone is impersonating Polson), hopefully he’s working on squashing this! (Also: what does the impersonator gain?)
Maybe they want to hurt/sabotage him specifically?
I dub this a “Polson process”
A “Polson Distribution”?
The answer is #2. Here is his real SSRN page (linked to from his UChicago home page): it lists 16 papers.
I guess there is still a possibility that he made the fake page himself as an experiment.
Interesting. The next question is, if someone is impersonating Polson, why? Some bizarre practical joke? Whatever the motivation, it seems likely that the papers were written as well as posted by a bot, no?
I have no idea what it means for someone to have two different SSRN pages.
Here’s the one with the fake papers:
Here’s the one with the real papers:
He’s Nicholas in the fake profile and Nick in the real one.
This was my finding too. After reading the post, I looked the guy up and quickly found the inconsistency in the two SSRN profiles. So, why didn’t Jeremy Horpedahl find this? What does he mean when he writes that “he just discovered an economist…?” No one checks these things anymore?
I dunno, it’s a mystery for sure why there are two profiles. This is consistent with the “social experiment” hypothesis and also the “impersonation” hypothesis.
“it seems likely that the papers were written as well as posted by a bot”
This. I liked Raghu’s theory, but then I asked Bing Copilot what was going on, and it agrees with Kaiser. No chance all this came from the same researcher. It offered possible explanations, the most like of which was “An AI‑assisted mass‑posting pipeline using [Polson’s] identity.”
I wrote this up here: https://www.junkcharts.com/what-can-one-believe/
Since it’s a statistics blog, we should note that the red flag is the citation to downloads ratio on the “fake” page.
My guess is that Polson’s having fun, or “an experiment” if you like. These are just working papers, so far. How many will get published? A very interesting question. What fields will publish AI papers? Can Polson create a record worthy of simultaneous chairs in sociology, English, and education? Can some of these papers get published in econ journals? Will they get cited?
The SSRN profile page linked from his webpage (https://papers.ssrn.com/sol3/cf_dev/AbsByAuth.cfm?per_id=22857) seems to be different from the one you posted. Supports the impersonation hypothesis.
FWIW: Polson has a (purported) 10 page proof of the Riemann Hypothesis posted to math.GM dating back to 2017, way before LLMs hit the scene (https://arxiv.org/abs/1708.02653) that he seems to still update (last revision this July) and take seriously; this personally moves my posterior quite a bit towards some amount of crankery amplified by LLM assistance, even though he seems to be a serious researcher and mathematically competent in other research areas. (Unless this is of course also just a prank, in which case he got me :))
I wonder if the papers have some nonobvious feature that we are missing. For instance, they might cite one another, with the whole project being an experiment on citation networks. Or perhaps they have some non-obvious watermark (e.g. some unlikely turn of phrase or something), and he is going to see if one can find evidence that they have been devoured by future AI models. This kind of thing might justify dumping 258 papers onto the internet.
His presentation of “the cumulative model” reminds me of that Ali G episode where Sacha Baron Cohen pitches his hoverboard idea to a potential investor.
From the post:
”Jesus Christ. I’ve heard that Cambridge University has an open position in their school of education . . . ”
This is an absolutely disgusting joke to make about someone who was hounded into suicide.
I agree. Please don’t do that.
I stumbled over this as well. Disgusting.
Anon:
I’m sad that this Cambridge University professor fabricated his research, I’m sad that Cambridge hired him without realizing the problem, I’m that they hired a law firm to intimidate the press from reporting about it and that they paid people to lie about, I’m sad that all this came tumbling down so quickly on the person, and I’m saddest of all that he killed himself. Sometimes I joke about things that make me sad.
Your “sadness” does not excuse you from treating others with basic human decency.
Anon:
I hope you’re posting comments on the Cambridge University website because I don’t think the Cambridge University executives treated their students with basic human decency when they hired an unqualified professor, I don’t think they treated those journalists with basic human decency when they sicced an evil law firm on them, I don’t think they treated their public relations staff with basic human decency when they instructed them to lie in public, and I don’t think they treated Jason Arday with basic human decency by treating him like some sort of publicity object and leaving him to hang out to dry rather than at least trying to walk him down from his fabrications. The well-paid executives at Cambridge University profited from a series of lies, and they were willing to sacrifice the education of their students and the integrity and welfare of their employees in a years-long attempt to keep the bubble from bursting. That makes me sad. And angry. I will also sometimes joke about such things. Joking about it doesn’t make me less sad or angry; it just provides me with another route into the problem.
Basic human decency also does not excuse you from callously dismissing Andrew’s sadness and his choice of how to express it.
What does “basic human decency” mean?
Is your thought policing treating people with basic human decency?
Who was hounded into suicide? Not Arday, as he decided to kill himself when the jig was up instead of being held accountable for his lies and fraud.
Are we sure it is the real Polson? His Chicago booth page links to another SSRN account with only 16 papers on it in total…
Booth page: https://www.chicagobooth.edu/faculty/directory/p/nicholas-polson
SSRN it links to: https://papers.ssrn.com/sol3/cf_dev/AbsByAuth.cfm?per_id=22857
Curious, I decided to google myself. Google responded with:
AI Overview
University of Chicago Booth School of Business professor Nicholas G. Polson has published an extraordinary volume of over 250 working papers on SSRN in 2026 alone.
Yep, that’s an AI overview, all right!
That was “google ssrn polson myself”. I put angled brackets around the search term and the blog comments editor went ahead and ate the search term.
This calls for more AI to fix the problem :)
Just a slightly more extreme case of what many AI/ML researchers are up to these days.
Polson has co-authors listed on many of the 256 papers. I am aware that at least one of these co-authors has requested that his name be removed as a co-author, which suggests to me that this person knows this is not an experiment. To add to the 6 hypotheses offered in explanation, I suggest another possibility: AI addiction at a scale heretofore unseen.
If that is the case, then this is far more serious than personal AI addiction, and verges on research misconduct. Nobody needs to “request” their names to be removed after posting to a preprint server. If any co-authors have reservations, the preprint should not be posted, to begin with.
Andrew,
I am Vadim Sokolov, coauthor of Theories of Human Connection and several other papers discussed in this post.
There is one substantive criticism in your post that I think is fair. Our discussion of Gottman repeats the reported claim that relationship outcomes could be predicted with accuracy above 90%. Gottman and Levenson reported 93% accuracy in a peer-reviewed paper, and related results have been published and widely discussed in the relationship literature. We were not aware at the time of the important methodological criticism concerning cross-validation and out-of-sample prediction. Having now reviewed that literature, we agree that our statement requires qualification, and we will correct it. But the rest of your post goes far beyond that and there a gap between the evidence you collected and the conclusions you published.
From one paper you speculate about a bot autonomously producing and uploading papers, authors not knowing what is in their own work, “intellectual squatting,” and even impersonation. You then estimate that the net scientific value of the entire collection is negative.
In your subsequent update you go further and refer to the SSRN page as containing “all the fake papers.”
That characterization is false. You are entirely entitled to think Theories of Human Connection is a bad paper or to think the “Master Equation” is silly. Neither establishes that the papers are fake.
In fact, one of these papers, Kramnik vs. Nakamura: A Chess Scandal (that you also mentioned), went through multiple rounds of referee review and revision before being published in CHANCE, a peer-reviewed American Statistical Association publication. It is a concrete example of why describing the entire collection as “fake papers” is simply not an evidence-based characterization.
These are genuine preprints produced and posted by the named researchers. The SSRN account was not an impersonation. Some of these papers may contain errors. They are preprints, and we certainly do not claim that hundreds of manuscripts produced across different subjects and at different stages of development are all error-free or equally good. But an erroneous or imperfect preprint is not a “fake paper.”
Nor did these projects suddenly materialize from a magical “chatbot” in 2026. Many grew from ideas, research notes, mathematical work, computational experiments, lecture material, and partial drafts developed over many years. Some are much newer and were developed with much more extensive AI assistance. Modern AI made it possible to develop and complete this material at a speed that would previously have been impossible. The collection is heterogeneous both in provenance and in the role AI played.
You have known Nick professionally for years. You could have asked him about those facts!
Indeed, later in your own post, when you briefly inspect “Bayes with No Shame”, you say that it looks “much more substantive” and “much more like a real scientific paper”, while also saying that you do not have the patience to read it. That should itself suggest the danger of treating hundreds of heterogeneous manuscripts as though they had already been evaluated.
The progression here is striking:
A claim in one preprint was inadequately qualified.
Therefore the paper is “bullshit.”
Therefore perhaps the authors did not read their papers.
Perhaps a bot generated and uploaded them autonomously.
Perhaps somebody was impersonating Nick.
Perhaps the entire corpus has negative value.
And then: “all the fake papers.”
Your blog is one of the most widely read in statistics. Statements by an influential scholar characterizing another researcher’s work as “fake” have consequences. Your own comment section already illustrates this: one commenter says that if Nick submits work to his journal, he now knows not to spend time on it. Shortly after these posts appeared, SSRN also removed our accounts and papers (although, might not be related to your post). But the sequence illustrates why precision matters when making serious public claims about other researchers. “I found a problem in this preprint” and “these are all fake papers” are fundamentally different claims.
Vadim Sokolov
Vadim:
Thanks for the info. Here’s the key bit I gather from your comment:
I guess that “Some are much newer and were developed with much more extensive AI assistance” applies to most of the 258 papers. It’s hard to know since I didn’t see anything in the papers noting the initial prompts, acknowledging the chatbot contributions, etc.
Given all this, I stand by my recommendation in the above post;
Regarding my labeling of the papers as “fake”: I do think that if you stick in some prompts, tell a chatbot to write a 30-page paper on the topic, and then label it as your own writing without acknowledgement, that’s fake. Nobody’s out there legitimately writing 258 papers in a single year. Someone can have 258 original ideas in a year—that’s no problem at all—and that’s why I suggest that the real contribution could be made by the single article listing and discussing the ideas.
I found it hard to believe that human would have read those 258 articles from beginning to end, even its named authors. What is that—10,000 words of computer-generated text? I guess it’s possible.
Finally, let me emphasize that:
1. I appreciate your commenting on this post and sharing your perspective.
2. I sincerely recommend you and your colleagues follow my suggestion and work out the ideas you find interesting and write them up in your own words, rather than producing thousands of pages of chatbot output which will be difficult for anyone to go through.
Quote from above: “Shortly after these posts appeared, SSRN also removed our accounts and papers (although, might not be related to your post).”
I assume SSRN has some rules concerning multiple accounts for one person, etc. which might explain why (some of?) the accounts were removed. If there was an original, older, account on SSRN from one or more researchers involved that might explain the actions of SSRN.
Multiple accounts for a single person may also help explain why people speculated about the possible fakeness of the new account, which in combination with the extraordinary amount of published papers might be considered to be a valid and useful thing to wonder about and consider.
I could also find the following in the terms of use of SSRN which might possibly be relevant here, or may help explain the actions of SSRN:
“Each registration is for a single individual only, unless specifically designated otherwise on the registration page. Elsevier does not permit a) anyone other than you to use the sections requiring registration by using your name and password; or b) access through a single name being made available to multiple users on a network or otherwise.”
Vadim:
Let me say just one more thing here. If you and Nick have ideas you want to share, that’s great. If you think that the best way to do this is to use chatbots to produce 258 papers with tens of thousands of pages of text, that’s your call—although I do think the responsible thing here would be to have a footnote on page 1 of each paper explaining the composition process, along with an appendix giving your prompts. I just don’t think that producing and posting this mass of computer-generated words will serve you well in your larger goals of exploring and communicating your ideas.
A few factual clarifications.
The workflow was not “write a prompt, ask a chatbot for a 30-page paper, and upload the result.” Many papers began with existing ideas, notes, mathematics, code, or drafts; others were newer. AI was then used iteratively in developing and writing them. The degree of AI involvement varied substantially across papers. Different papers had different histories and different levels of involvement by individual authors; there was no single uniform production process across the corpus.
None of this justifies describing the corpus as “all fake papers.” Nor does it establish that the papers resulted from the one-prompt/one-paper workflow described in these posts.
On disclosure: the extent of AI assistance was not consistently disclosed, and I agree that for papers with substantial AI involvement it should have been disclosed.
On the Riemann hypothesis: I have no involvement in that work and cannot speak to it. Characterizing a colleague as “delusional” and combining that with my interest in AI to explain “hundreds of junk papers” is speculation about people rather than analysis of the papers.
Why were the papers deleted after Andrew’s post came out?
Vadim:
Thanks for the clarifications.
Regarding the Riemann hypothesis: I don’t know about the math, but given the reputation of this conjecture, I’d guess that if some statistician thinks he’s proved it and nobody else thinks that, then he’s indeed delusional. It’s a speculation in the same way that if some amateur athlete told me he ran a 3:30 mile, I’d suspect he’s either delusional or faking it. And I could well believe that the same delusion that would make Nick think he has proved the Riemann hypothesis could also lead him to think that these AI-produced papers have genuine mathematical value.
Again, I encourage all of you to work out your ideas and write them out directly without producing tens of thousands of pages of computer-generated words. Ultimately the goal should be to work out the ideas and to communicate them to people, not simply to pile up mountains of words that nobody will read and often make no sense at all.
I can’t carefully _read_ 6 or more papers in a day, let alone discern if I should believe them and figure out how they might apply to my related questions of interest. Can the audience be anything other than bots & AI ingestion?
I’m more interested in how this form of “mass production” will affect SSRN and other preprint services.
1) SSRN claims to ensure the posting of high-quality research not by peer-review, but by research integrity filters, including AI usage with no disclosure, AI and AI technologies as an author, and paper mill contributed content. I suggest they need a better filter. https://www.elsevier.support/ssrn/answer/research-integrity
1a) Perhaps their filters are good enough for this case but simply not fast enough, and as they work through the backlog of Polson preprints they flag and remove ach one, leading to the gradual shrinking number of preprints on the Polson page?
2) Will this mass production inflate the haystack of preprints enough that the ability to find relevant and even marginally worth skimming preprints is lost, and preprint services die out as n longer useful?
3) Will this mass production overwhelm the possibility of thoughtful (or any human) peer review? Or will there need to be some submission fee that goes toward peer reviewers, or some rule that after a couple of papers published someone cannot even coauthor a submitted paper unless they provide non-superficial peer reviews on some number of other submissions and keep their review to submission ratio high enough?