Eamon Duede
@eduede.bsky.social
epistemology of science and artificial intelligence // asst prof purdue university // argonne national lab // phd, university of chicago www.eamonduede.com
Nature did a nice, accessible write up of a recent piece I wrote with @mjcrockett.bsky.social, Kevin Gross, and @carlbergstrom.com. Also, check out Carl's more detailed thread (with link to the paper in both). "Scientists using LLMs will ‘do more, less well'" www.nature.com/articles/d41...
Scientists using LLMs will ‘do more, less well’, modelling study predicts
Research suggests that the combination of incentives to publish and the use of large language models will lead to more papers, but they will be less refined.
nature.com
Carl T. Bergstrom@carlbergstrom.com · 2w ago1. We—@eduede.bsky.social, @mjcrockett.bsky.social, Kevin Gross, and I—have a new preprint on the arXiv today, based on ideas that emerged during an @sfiscience.bsky.social workshop in November 2024: The unintended consequences of large language models as a labor-augmenting technology in science.
"Hence our dismal conclusion: rather than finding that LLMs free us to do a better job of what we were doing before they came along, they shift scientific incentives (and the playing field of academic competition) in ways that compel us to do more and more, faster and faster, less and less well."
Carl T. Bergstrom@carlbergstrom.com · 2w ago1. We—@eduede.bsky.social, @mjcrockett.bsky.social, Kevin Gross, and I—have a new preprint on the arXiv today, based on ideas that emerged during an @sfiscience.bsky.social workshop in November 2024: The unintended consequences of large language models as a labor-augmenting technology in science.
New preprint with @carlbergstrom.com, Kevin Gross, and @mjcrockett.bsky.social. See Carl's thread. Surprise! Even if LLMs work perfectly, they'll make scientists refine their work less deeply, because they raise the opportunity cost of time, pushing researchers to do more, less well.
Carl T. Bergstrom@carlbergstrom.com · 2w ago1. We—@eduede.bsky.social, @mjcrockett.bsky.social, Kevin Gross, and I—have a new preprint on the arXiv today, based on ideas that emerged during an @sfiscience.bsky.social workshop in November 2024: The unintended consequences of large language models as a labor-augmenting technology in science.
If you're at #ICLR and interested in #humanities centered work on #AI (plus a radical take on the problem of #LLM hallucination), check out our paper arxiv.org/abs/2511.07722 on Fri, Apr 24 at 11:15 AM – 1:45 PM PDT Pavilion 3 P3-#1819. Joint work w/ @richardjeanso.bsky.social @hoytlong.bsky.social
Critical Confabulation: Can LLMs Hallucinate for Social Good?
LLMs hallucinate, yet some confabulations can have social affordances if carefully bounded. We propose critical confabulation (inspired by critical fabulation from literary and social theory), the use...
arxiv.org
As AI starts proving theorems, a question becomes urgent: when does a formal derivation correspond to a particular proof idea—not just prove the same theorem a different way? @simondedeo.bsky.social and I argue this 'correspondence problem' has been hiding in plain sight. arxiv.org/abs/2603.13680
A correspondence problem for mathematical proof
Mathematical proofs are often said to justify their conclusions by indicating the existence of a corresponding formal derivation. We argue that this widespread view relies on an under-examined notion ...
arxiv.org
The talk is based on a paper I co wrote with @eduede.bsky.social @hoytlong.bsky.social & Patrick Sui that is accepted and forthcoming at ICLR 2026. Here's the preprint: arxiv.org/abs/2511.07722
Critical Confabulation: Can LLMs Hallucinate for Social Good?
LLMs hallucinate, yet some confabulations can have social affordances if carefully bounded. We propose critical confabulation (inspired by critical fabulation from literary and social theory), the use...
arxiv.org
The schedule and RSVP link for the CAIC launch event is now live on our website 🚀 Check the schedule and RSVP to attend ✏️ caic.centers.purdue.edu
Cool postdoc: hiring someone with training in philosophy + computational methods to study how AI is transforming epistemic norms & practices across academic fields of knowledge. Can vouch the PI is both a fantastic advisor & person. Great academic network too: Purdue to CMU to Chicago to Princeton.
Eamon Duede@eduede.bsky.social · 6mo agoExcited to recruit a 🚨 postdoc 🚨 for projects on #AI and evolving scientific practice, norms, and impact! Interdisciplinary work across science of science, philosophy of science, math with creative Purdue/Argonne/CMU/Chicago/Princeton collaborators! *pls repost*! careers.purdue.edu/job/Postdoc-...
New paper on Why Slop Matters w/ great group of co authors (@hoytlong.bsky.social @eduede.bsky.social @ari-holtzman.bsky.social + others not on Bluesky) from ACM AI Letters. We try to move the debate re: AI Slop past normative, neg claims & towards parsing its social uses. dl.acm.org/doi/10.1145/...
Why Slop Matters | ACM AI Letters
AI-generated “slop” is often seen as digital pollution. We argue that this dismissal of the topic risks missing important aspects of AI Slop which deserve rigorous study. AI Slop serves a social funct...
dl.acm.org
Excited to recruit a 🚨 postdoc 🚨 for projects on #AI and evolving scientific practice, norms, and impact! Interdisciplinary work across science of science, philosophy of science, math with creative Purdue/Argonne/CMU/Chicago/Princeton collaborators! *pls repost*! careers.purdue.edu/job/Postdoc-...
Postdoc Research Assoc/Sloan Fdn
Postdoc Research Assoc/Sloan Fdn
careers.purdue.edu
New on the Archive: Avigad, Jeremy (2026) Mathematical Understanding. [Preprint] https://philsci-archive.pitt.edu/27708/
New on the Archive: Duede, Eamon and Friedman, Daniel (2025) Epistemic Gaps and the Attribution of (AI) Discovery. [Preprint] https://philsci-archive.pitt.edu/27719/
Excited this piece is finally out in Philosophy of Science. We argue that, paradoxically, we can have certainty about theorems in math, the proofs of which we can never understand. That's weird and tells us something important about why we do math in the first place #philsky tinyurl.com/3dxxb98a
A Priori Knowledge in an Era of Computational Opacity: The Role of Artificial Intelligence in Mathematical Discovery | Philosophy of Science | Cambridge Core
A Priori Knowledge in an Era of Computational Opacity: The Role of Artificial Intelligence in Mathematical Discovery
cambridge.org
Reviewer 2 may have used the autopen during the final days of the Biden administration
Can AI simulations of human research participants advance cognitive science? In @cp-trendscognsci.bsky.social, @lmesseri.bsky.social & I analyze this vision. We show how “AI Surrogates” entrench practices that limit the generalizability of cognitive science while aspiring to do the opposite. 1/
AI Surrogates and illusions of generalizability in cognitive science
Recent advances in artificial intelligence (AI) have generated enthusiasm for using AI simulations of human research participants to generate new know…
sciencedirect.com
1. We ( @jbakcoleman.bsky.social, @cailinmeister.bsky.social, @jevinwest.bsky.social, and I) have a new preprint up on the arXiv. There we explore how social media companies and other online information technology firms are able to manipulate scientific research about the effects of their products.
New piece w/ James Evans in Science explores what we call 'science after science', an era where our ability to control nature may exceed our ability to understand it; a new struggle to sustain curiosity & understanding under AI's predictive dominance. #ai #science www.science.org/doi/10.1126/...
After science
Twenty-five years ago, Ted Chiang wrote a prescient science fiction short that began: “It has been 25 years since a report of original research was last submitted to our editors for publication, makin...
science.org
Twin cities friends: I'll be speaking tomorrow at the Minnesota Center for Philosophy of Science on how AI disrupts the sometimes precarious balance of scientific incentives.
Minnesota Center for Philosophy of Science@mcps-philsci.bsky.social · 9mo agoExcited to welcome @lmesseri.bsky.social, @cameronbuckner.bsky.social and @carlbergstrom.com to campus a week from today for the “AI and the Nature of Science: Concepts and Controversies” event. 14.11.25, 1400 - 1800.
One of the best #philsci conferences in one of the best cities: the 2026 BSPS Annual Conference is taking place in Leeds 21–23 July. Keynotes by Robin Hendry, Tarja Knuuttila & Eleanor Knox. More info: www.thebsps.org/news/bsps-an... #philsky
BSPS Annual Conference 2026
The BSPS Annual Conference takes place on 21–23 July 2026 at the University of Leeds
thebsps.org
it’s crazy because no one knows what the aim of science is
In 1977 computers proved the 4Color Theorem. Human's can't check that proof, but trust it because they understand it. Now #AI can generate proofs we'll never understand. Can we trust those? In a new paper out in PhilSci @philscijournal.bsky.social, we argue yes! with a catch: tinyurl.com/yhmnrx5m
Apriori Knowledge in an Era of Computational Opacity: The Role of AI in Mathematical Discovery | Philosophy of Science | Cambridge Core
Apriori Knowledge in an Era of Computational Opacity: The Role of AI in Mathematical Discovery
cambridge.org
1. The philosophy of science sometimes gets an unearned reputation as a purely academic exercise that offers little by way of concrete tools for advancing research. This is wrong. And today, as we grapple with how AI is changing the nature of scientific activity, it's desperately wrong.
Applications are open for SFI's 2026 Complexity Postdoctoral Fellowships If you’ve recently earned a Ph.D. in any scientific field and want to pursue independent, transdisciplinary research, consider applying. Deadline: October 1, 2025 Apply here: santafe.edu/sfifellowship
A few months ago, Nature published how-to guide for using ChatGPT to write your peer reviews in 30 minutes. This is, of course, a horrible idea. Here’s my response with @jbakcoleman.bsky.social .
AI, peer review and the human activity of science
When researchers cede their scientific judgement to machines, we lose something important.
nature.com