The Philosophers' Cocoon

A safe and supportive forum for philosophers.

Owned & Moderated by Marcus Arvan (University of Tampa)

Contact: marvan@ut.edu

Blog mission & moderation policy

IMPORTANT NOTE on how to ensure anonymity in commenting
Job-Market Reporting Thread (2026-27)
Job-Market Discussion Thread (2026-27)

recent posts

about

Are unpublished preprints degrading the quality of work in philosophy of AI (and other fields)?

In our October “how can we help you?” thread, a reader writes:

The recent discussions on the state of Philpeople/PhilPapers has got me thinking about the general role of preprints on sites like arXiv in the currently literature, particularly in ai-related work.

Not entirely sure to what extent this is a new phenomenon, but my impression is that philosophy of AI is following computer science in making heavy use (heavier than the rest of philosophy, at least) of such sites; papers are posted before they’ve undergone any kind of peer review, apparently in a race to say something new and start gathering citations as soon as possible.

For sure, if a paper is of high quality, there’s some value in people being allowed to read it as soon as possible. But I do think this practice leads to some low quality stuff being posted, and it might encourage authors to “dig in their heels” and avoid making significant changes to their views or to the papers when they do eventually get pee reviewed. And even when the work is bad for these reasons (we could call it “fast philosophy”), there’s pressure to cite it anyway in order to ensure one’s references are up to date.

I don’t post any of my work as preprints, and I feel weird about citing some of this stuff, because it feels like rewarding it.

I’d be interested to hear others thoughts about this. Do others also think preprints are degrading the quality of work in philosophy of AI (or other subfields)? Should philosophers post more of their work as preprints before it’s been peer-reviewed? Less? Should we be citing it more or less?

These are really good questions. What does everyone think?

Posted in , ,

3 responses to “Are unpublished preprints degrading the quality of work in philosophy of AI (and other fields)?”

  1. Federico Faroldi

    I closely follow fields where the arXiv is used a lot. You see an invasion of “philosophical” work coming out e.g. of the labs that is, on average, of very poor philosophical quality. My impression is that those authors want to engage with and influence other communities (e.g. CS, the other labs, funding orgs), rather than with professional philosophers — a different ballgame. It is true, however, that the philosophy journals are extremely slow; often, a paper is out of date (think of system capabilities, agentic architectures, etc.) if one has to wait a year or two. Personally, I don’t post philosophical work on the arXiv either.

    Reply
  2. sahpa

    Obviously none of this is fully separable from the broader, ongoing debate about philosophy’s/academia’s reliance on peer review, so suffice to say that we are by no means agreed that peer-reviewed journal publications should be the professional standard.

    As for philosophy of AI in particular, I take it that subfield has a few specific arguments for the arXiv model:

    • computer scientists and such are among their key intended readerships, and they read preprints as a matter of course
    • the field of AI is moving much more rapidly than the philosophy journal lifecycle can handle, so peer-reviewed publications are often/normally stale before they come out
    • papers are often co-written with technical people with a different set of disciplinary norms and expectations
    • there can be career benefits to following the preprint/arXiv model for all of the above reasons that extend well outside the philosophy profession itself — to interdisciplinary research institutes, think tanks, AI labs, etc.

    And that’s true even if you are bullish about peer review — thinking, as OP suggests, that it is an important mechanism for increasing quality, reducing stubbornness, and reducing the incentive to cite bad papers just to ensure your bibliography gets past the reviewers and is ‘up to date’.

    Reply
  3. Anonymous

    I’d encourage you to look into the literature on the effectiveness of peer review. There are substantial, well-documented problems with it, so I don’t think we should assume that peer-reviewed work is necessarily of higher quality. On the other hand, I’m not aware of any evidence supporting the concerns you’re raising about preprints (e.g., that posting a preprint makes authors more likely to “dig in their heels”).

    More generally, if someone cites a paper, it should be because they’ve read it, evaluated its arguments, and decided it’s worth citing—not because it has passed peer review. Peer review is not a substitute for exercising one’s own scholarly judgment. It would be backwards to discourage citing a paper simply because it hasn’t yet undergone peer review.

    Reply

Leave a Reply