AI-generated research papers are being published everywhere

Originally published at:



This is perhaps the biggest near-term threat from AI. Not that it will try to kill us or even take our jobs, but that it can be used to create mountains of garbage that will make it nearly impossible for most people to sort out good information from bad.


And increasing over the Summer months when academics are on vacation.

Enter. Another margarita please.


Not only that, but future LLM models will be trained on the mountains of garbage, leading to what Jathan Sadowski calls “Habsburg AI”:

a system that is so heavily trained on the outputs of other generative AI’s that it becomes an inbred mutant, likely with exaggerated, grotesque features.

That’s before you get to Bonesaw AI.


The AI apocalypse isn’t going to be skynet or some other rogue AI launching nukes… No sir, we’re just going to keep pushing pollution via the tech sector and we’re all gonna drown in this kind of BS…

See Jon Ossoff GIF by Election 2020

Frustrated Beth Behrs GIF by CBS

The Muppets No GIF by ABC Network

It was a right wing attack on academia is what that shit was.

Could it be because we have ever fewer tenure-track and tenured professors who have the TIME to do the UNPAID labor of running these journals? So many of us are barely scrapping by as adjuncts, that we don’t have time to also edit or do peer reviewed work like this. Some of us have been calling this shit out for ages, only to be admonished and told to shut up and do better work if we want a better position. And now the far right are coming for tenured jobs altogether.


It seems incredible to me that damn near everyone pushing “AI” so fucking hard seems to neglect factoring in the ecological cost. Just like crypto-currency, that shit consumes hella energy, and none of it is sustainable, especially in the face of grid-fail.


Honestly, it’s a problem with the entire tech sector. Think about all those data centers, sucking up power in cities across the country… I’m currently reading Blood in the Machine by Brian Merchant, and the similarities between the emergence of the industrial economy at the beginning of the 19th and the tech sector today is quite striking (including things like ignoring environmental costs and the cost to the workers who got screwed over by the rise of industrial labor).


I swear, if society ends being on Fury Road and I am one of the survivors, I will make it my life’s mission to constantly remind the rest who it was that “killed the world.”


But we also found AI-generated language in more reputable academic journals. Last week, Bellingcat researcher Kolina Koltai tweeted a paper published in Surfaces and Interfaces, which included the phrase “Certainly, here is a possible introduction for your topic:” at the top of the paper’s introduction.

But is “Surfaces and Interfaces” a reputable academic journal?

Here’s a page listing an Article publishing charge for open access of $2,360.

Here are graphs showing average review times below 6 weeks, followed by an average online publication time less than a week. The first time is short by industry standards, and the second is so short that I suspect they routinely start on that job before they have the review result.

Here’s an article that’s been online since 2019 accusing Elsevier of publishing predatory journals, an accusation that surprised hardly anyone.

So, lots of red flags for this journal. Is this the best example that 404 Media could find of a “more reputable journal” publishing dodgy papers?

And I’ve don’t have the skills or knowledge to be able to read papers in the field of Surfaces and Interfaces to decide if they are genuine scholarship. All I can do is look for the generic warning signs. If anyone with appropriate Engineering skills wants to tell me that this is actually a good journal, I’m happy to listen.


Agreed. And hardly worth the hyperbolic headline for this post. “Everywhere”?

Nevertheless… Surfaces and Interfaces today, maybe PLoS One next year? The fact that it’s happening at all should raise alarm.


This topic was automatically closed after 5 days. New replies are no longer allowed.