AI & Cognitive Work. Cambridge University Press. 2026.

AI will either displace or assist human cognitive workers; displacement is much less bad than most people think, and assistance is far worse.

Epistemic Instrumentalism Explained. Routledge. 2023.

People have a reason to believe something when and because doing so would promote their interests.

1. Cognitive Enfeeblement By Design. Synthese. Forthcoming.

   Developers at large AI labs are deliberately designing tools that mechanically reduce both the opportunity and the incentive to engage in cognitively difficult tasks.

2. Language Models for Value-Informed Proxy Decision Support. (With Victoria Nolan, Marcia Brown, Jeremy Balch, et al.) New England Journal of Medicine AI. 2026.

   LLMs could be useful for proxy decision support, but currently do things like predict preferences based on patients' perceived literacy, which most people don't like.

3. Extreme Cognitive Assistance and Open Futures. Journal of Ethics and Social Philosophy. 2026.

   Gemini, Claude, and other frontier AI assistants will cognitively enfeeble kids, and this is morally bad.

4. Promotionalism, Orthogonality, and Instrumental Convergence. Philosophical Studies. 182 (7):1725–1755. 2025.

   Even extremely capable AI probably won't kill us as a means to some other end like making lots of paperclips.

5. What is it for a Machine Learning Model to Have a Capability? (With Jacqueline Harding.) British Journal for the Philosophy of Science. 2025.

   Frontier machine learning systems can do a lot of things, but it is unclear what it means to say this.

6. Personalized Patient Preference Predictors are Neither Technically Feasible Nor Ethically Desirable. American Journal of Bioethics. 24 (7):62–65. 2024.

   Fine-tuning LLMs on social media data to predict people's preferences regarding their end-of-life care is a bad idea.

7. Morality First? AI & Society. 40 (3):1289–1301. 2025.

   Focusing post-training in large-scale AI systems on "ethics" and "morality" risks misrepresenting the normative landscape.

8. Predicting and Preferring. Inquiry: An Interdisciplinary Journal of Philosophy. 68 (4):1121–1132. 2025.

   Capable ML systems trained to predict patient preferences have a perverse incentive to try and make patient preferences more predictable.

9. Spot the Bot. AI & Society. 40 (2):1113–1115. 2025.

   It's not possible to reliably identify artificial agents in the wild, and large companies know this.

10. Large Language Models and Biorisk. (With Harry Lloyd and William D'Alessandro.) American Journal of Bioethics. 23 (10):115–118. 2023.

    LLMs may make bio-terrorism much easier, and this is very bad.

11. Is Deontic Evaluation Capable of Doing What it is For? (With Rob van Someren Greve.) Journal of Ethics and Social Philosophy. 19 (3). 2021.

    Evaluation in terms of "right/wrong" and "obligatory/permissible" is not necessarily capable of guiding action just because we would prefer that it be.

12. Ecumenical Epistemic Instrumentalism. Synthese. 198: 2613–2639. 2021.

    Epistemic instrumentalists can be more ecumenical by accepting a conditional rather than an existential claim.

13. Should Aggregate Patient Preference Data Be Used to Make Decisions on Behalf of Unrepresented Patients? AMA Journal of Ethics. 21 (7):566–574. 2019.

    Using statistical models to predict patients' preferences faces significant moral barriers to implementation.

14. Promotion as Contrastive Increase in Expected Fit. (With Finnur Dellsén.) Philosophical Studies. 176 (5):1263–1290. 2019.

    Promoting a desire is a matter of increasing someone's expectation that the world more closely fits that desire.

15. Consequentialism & Moral Worth. Utilitas. 31 (2):117–136. 2019.

    Consequentialists have no special problem accounting for the "moral worth" of actions or the "right reasons" to do things.

16. Epistemic Instrumentalism and the Reason to Believe in Accord with the Evidence. Synthese. 195 (9):3791–3809. 2018.

    Epistemic instrumentalists should explain away the intuition that there's always a reason to believe in accordance with one's evidence.

17. Two Problems for Accepting as Intending. Ethics. 128 (3):626–641. 2018.

    Accepting someone's promise to make you breakfast doesn't give you any reason to help them make you breakfast.

18. Patient Preference Predictors and the Problem of Naked Statistical Evidence. Journal of Medical Ethics. 44 (12):857–862. 2018.

    Predicting whether someone wants to be resuscitated on the basis of their race is a lot like predicting they will commit a crime on that basis.

19. Rational Coherence in Environmental Policy: Paris, Montreal, and Kigali. Ethics, Policy and Environment. 21 (1):4–8. 2018.

    Manufacturing reasons to support climate action is probably better than ignoring the good reasons to do so.

20. The Beliefs & Intentions of Buridan's Ass. (With Finnur Dellsén.) Journal of the American Philosophical Association. 3 (2):209–226. 2017.

    If you're going to intend to eat one bale of hay rather than another identical bale, you had better have sufficient reason.

21. In Defense of Comic Pluralism. Ethical Theory and Moral Practice. 20 (2):375–392. 2017.

    The moral qualities of a joke can both increase and decrease its comedic qualities.

22. A Partial Defense of Permissivism. Ratio. 30 (1):57–71. 2017.

    Roger White's arguments against the permissiveness of epistemic rationality are not persuasive.

23. Checking the Neighborhood: A Reply to DiPaolo & Behrends on Promotion. Journal of Ethics and Social Philosophy. 2016.

    The idea that promoting a desire involves making its satisfaction more likely is still wrong.

24. Fairness and the Strengths of Agents' Claims. Utilitas. 28 (3):347–360. 2016.

    Distributing indivisible goods fairly is tricky, but we should probably use a lottery.

25. How You Can Reasonably Form Expectations When You're Expecting. Res Philosophica. 92 (2):1–12. 2015.

    L.A. Paul is wrong that it is structurally impossible to reasonably decide not to have children.

26. On Durant Drake's 'May Belief Outstrip Evidence?' Ethics. 125 (2):536–539. 2015.

    Writing at the turn of the 20th century, Durant Drake was concerned about a topic of much interest to contemporary epistemology.

27. Problems for Pure Probabilism About Promotion (and a Disjunctive Alternative). Philosophical Studies. 172 (5):1371–1386. 2015.

    Satisfying some desires is structurally impossible, but it is still possible to promote them.

28. Reasons and Promotion. Philosophical Issues: A Supplement to Noûs. 25 (1):98–122. 2015.

    The relationship between normative reasons and promotion is fraught.

29. Reasons Wrong and Right. Pacific Philosophical Quarterly. 97 (2):371–399. 2015.

    Doing something for the "wrong reasons" means doing it for a special class of reasons that share a common feature.

30. Nothing but the Evidential Considerations? Australasian Journal of Philosophy. 94 (2):1–19. 2015.

    It can sometimes be rational to believe things for reasons that have nothing to do with one's evidence.

31. Schroeder on the Wrong Kind of Reasons Problem for Attitudes. Journal of Ethics and Social Philosophy. 7:1–8. 2013.

    Mark Schroeder's two solutions to the WKR Problem do not work.

• Depression and Suicide. In The Oxford Handbook of Suicide, eds. Michael Cholbi & Paolo Stellino. Oxford University Press. 2025.
• Epistemic Consequentialism: Haters Gonna Hate. In Metaepistemology – Realism and Anti-Realism, eds. Robin McKenna & Christos Kyriacou. Palgrave Macmillan, 2018.

• The Range of Reasons: In Ethics and Epistemology by Daniel Whiting (Oxford University Press, 2022). Ethics. 133 (3):461–465. 2023.
• Epistemic Consequentialism, eds. Kristoffer Ahlstrom-Vij & Jeffrey Dunn (Oxford University Press, 2018). Utilitas. 32 (2):256–260. 2020.
• Motivational Internalism, eds. Gunnar Björnsson et al. (Oxford University Press, 2015). Notre Dame Philosophical Reviews. 2015.