ArXiv will ban authors who submit papers with hallucinated citations

droidjj1 pts0 comments

Thomas G. Dietterich (@tdietterich): "Attention @arxiv authors: Our Code of Conduct states that by signing your name as an author of a paper, each author takes full responsibility for all its contents, irrespective of how the contents were generated. 1/" | XCancel

Thomas G. Dietterich@tdietterich

May 14

Attention @arxiv authors: Our Code of Conduct states that by signing your name as an author of a paper, each author takes full responsibility for all its contents, irrespective of how the contents were generated. 1/

May 14, 2026 · 7:03 PM UTC

86

692

4,163

623,215

Thomas G. Dietterich@tdietterich

May 14

If generative AI tools generate inappropriate language, plagiarized content, biased content, errors, mistakes, incorrect references, or misleading content, and that output is included in scientific works, it is the responsibility of the author(s). 2/

13

53

1,003

63,097

Thomas G. Dietterich@tdietterich

May 14

We have recently clarified our penalties for this. If a submission contains incontrovertible evidence that the authors did not check the results of LLM generation, this means we can't trust anything in the paper. 3/

53

990

96,774

Thomas G. Dietterich@tdietterich

May 14

The penalty is a 1-year ban from arXiv followed by the requirement that subsequent arXiv submissions must first be accepted at a reputable peer-reviewed venue. 4/

27

127

1,645

207,132

Thomas G. Dietterich@tdietterich

May 14

Examples of incontrovertible evidence: hallucinated references, meta-comments from the LLM ("here is a 200 word summary; would you like me to make any changes?"; "the data in this table is illustrative, fill it in with the real numbers from your experiments") end/

24

51

1,164

67,403

Dr. Tuğrulcan Elmas@DrTugrulcan

3h

Replying to @tdietterich @arxiv

I appreciate better enforcement then. A few months ago I sent a complaint about an AI paper with hallucinated references farming citations by squatting on a popular keyword and this is the brilliant response I got

1,304

Thomas G. Dietterich@tdietterich

3h

Did you file a Code of Conduct complaint with @arxiv?

903

Justin Angel

@JustinAngel

22h

Replying to @tdietterich @arxiv

My guess is that this policy will be applied selectively depending on institutional privilege and personal notoriety. It'll end up as a tool of silencing unconnected individuals vs. promoting better scientific discourse.

I aspire to be wrong.

12

140

21,536

Thomas G. Dietterich@tdietterich

22h

I agree that there could be biases in our pipeline. We apply a standard LLM detection algorithm to identify papers that need scrutiny. Moderators may also be biased. We would love to collaborate with researchers to study the bias and effectiveness of our operations!

10

210

18,956

more replies

Alessandro Melchiorri@alemelk

14m

Replying to @tdietterich @arxiv

16

Daniel Tan

@danieltanfh95

2h

Replying to @tdietterich @arxiv

This has to be a joke considering that arxiv uses AI tools to classify papers (in some cases wrongly as well but apparently nothing was done because…?) blog.arxiv.org/2020/05/20/ne…

400

MadMoxxie@yeahnah279

2h

Replying to @tdietterich @arxiv

This appears targeted. The need for assurance in the integrity of research data and references is crucial, but the quarantine of modern language tools is not. Seens like the problem is people of previous Gens dont like modern tools they didn't get to use . Change=Bad 🙄 🤦🏽‍♀️

258

Doctor Magdaki@Doctor6969420

1h

Replying to @tdietterich @arxiv

Glad to hear this. The quality of papers on arxiv has been steadily declining, reducing its value.

146

Ein Mensch@EinMensch2000

1h

Replying to @tdietterich @arxiv

Arxiv regularly rejected preprints of physics papers published in reputable physics journals long before AI existed.

Arxiv has long lost its reputation of being objective.

Arxiv will end like Wikipedia: it will only publish what fits the narrative.

168

Pierre Colmez@ColmezPierre

6h

Replying to @tdietterich @arxiv

That will be a lot of fun in domains where papers have an infinite numbers of co-authors...

330

Felipe Herrera@faherreraur

9h

Replying to @tdietterich @arxiv

This is positive for @arxiv, but some type of coauthorship confirmation step would need to be implemented to prevent misuse.

892

Bruce Lambert

@bruce_lambert

May 14

Replying to @tdietterich @arxiv

I think this is a perfectly reasonable and defensible strategy. I commend you and your colleagues.

155

14,641

Lucas Beyer (bl16)

@giffmana

May 14

Replying to @tdietterich @arxiv

Very good, please strongly enforce it.

219

13,773

Alex Kontorovich

@AlexKontorovich

22h

Replying to @tdietterich @LaboratoryMinds @arxiv

How do you tell whether authors agreed to be listed? If person X writes an AI generated slop paper, puts my name on it without my knowledge, and uploads it to the arxiv, am I banned too? Conversely, if person X and I write slop AI and he submits it to the arxiv and gets banned, can I...

arxiv tdietterich replying thomas dietterich authors

Related Articles