The reactions of many researchers on finally being held responsible for having read the very paper they submitted are... something.

The funniest thing about #arxivgate is that it’s going to drive everyone to use AIs to ensure their papers are in compliance because AIs are literally the only entity with the patience to read an entire paper.

if you get caught submitting AI slop to arxiv, the punishment should be generational aura loss

Although I generally feel something should be done about arXiv, I can understand the backlash given the severity of the proposed punishment. Could we at least institute a points-based system, like for a driving licence? A direct one-year ban, plus losing the ability to share unpublished work on arXiv, seems too harsh.

That said, failing to notice that a reference is incorrect is different from submitting a paper that contains an AI prompt, which could literally be detected by reading the said paper.

My issue with "you just need to read the paper" crowd: if I upload a paper with "TODO: Tom, please edit" in a caption nothing happens; so it feels weird to asymmetrically punish a mistake because AI was involved (and yes, mistakes can happen).

i see many variations of this claim. my read is that the premise is false. arxiv (rightfully) care about fighting AI slop. hallucinated citations just happen to be very easy to spot and very hard to dispute, so an easy target.

I strongly agree with the @arxiv ban on hallucinated references but it makes me wonder if papers are even the right format to disseminate knowledge now? If it is LLMs all the way down: writing, coding, reviewing, citing then we've automated a wrapper around the science. A narrative makes sense when you have one, not when it's forced. Maybe it's time to automate the verification instead. Agentic harnesses where you submit code and data, not a story?

@DellAnnaLuca I'm convinced only about 10% of college grads and grad students today would have passed and gotten their degrees 40 years ago.

@SianGooding @arxiv Glad that we are thinking about the same thing! Strip out the storying telling from research paper! We prose agent native research artifact, from conceptual level claims, implementation, ground truth results and to exploration trajectory

https://arxiv.org/abs/2604.24658

The easiest way to avoid the arxivgate: use AI. Agents are good at this.

It will force people against AI to use AI to check their papers out of fear as human make mistakes, and then they might fall in love with AI, which would be funny.

I think the 1-year ban is already quite a severe punishment. The life sentence of allowing to only post post-prints seems a bit extreme.

That said, I am all for testing (dis)incentives. I think @NeurIPSConf did not handle its issue with hallucinated references well.

"Dad, why do the other kids say I have no rizz?"

"Well son, let me tell you about a mistake your grandfather made decades ago with a model called Haiku 4.5...."

@CurmudgeonlyOpa Raising graduation rates beyond a certain point was mistake

@DellAnnaLuca the guy horrified that an author has to check his own citations and be able to read them if in a different language slays me.

This is kind of a shitpost but unironically, reputations (good and bad) should be well known and have some persistence. This is especially true given AI tools.

I don't think that my position is very radical. I agree that unchecked AI used is bad and fabricated citations are serious. I am mostly very afraid of the social consequences of this policy as I think many good scholars will end up being banned for minor offenses, purely by a large number effect

@DellAnnaLuca How is this even remotely controversial? You only put citations in a paper you've read, yes.

Yes, putting fabricated references at the same level of LLM meta-comments is very alarming

👀👀

This is such an ironic move by arxiv - exactly when codex can actually check your bibtex better than yourself.