Patrick H. Pan
I am a Ph.D. candidate in philosophy at UC Berkeley working mostly on the ethics of AI, normative ethics (especially contractualism), and moral psychology. My dissertation, advised by R. Jay Wallace and Johann Frick, asks whether it is ever fair to hold an artificial agent morally responsible for what it does, drawing both on the philosophical literature on responsibility in humans and on recent empirical work on LLMs. I also have an interest in Ancient Greek and Chinese philosophy and language.
Before Berkeley, I studied philosophy at Cambridge and computer science and piano in the Harvard–NEC dual-degree program, and worked as a software engineer.
Philosophy
all philosophy →- Ex Ante Contractualism and Fairness Individuals often accept small chances of serious harm for a certain modest benefit, and policymakers often impose the same bet on large populations. What makes this morally right or fair?
- The Problem of Double Vision: Subjective and Objective Perspectives We can see our beliefs and decisions as the product of our own agency, or as the predictable effects of external causes. Can we hold both views at once?
- Attention, Agency, Morality Iris Murdoch built her ethics around attention rather than right action. Where does attention fit if morality is about what we owe each other?
- The Disturbance Problem: How Predictive AI Models Reshape Preferences A model trained on dating-app swipes finds that the tilt of the eyes predicts attractiveness. What goes wrong when people start looking for it?
- “Deepfakes” and the End of the Photographic Age Computers can now make images that look just like photographs. What happens to photographs—either as evidence or as art—once no one can tell the difference?
- Dear Mr. Muskrat: or, How to Defeat Bad Faith Appeals to Non-Identity Mr. Muskrat argues that destroying the Amazon rainforest to build a racetrack will make no future person worse off, so it is permissible. What would it take to change his mind?
Elsewhere, recently
- 25 Oct 2026 Payare, Orchestre Symphonique de Montréal, Kavakos — Tchaikovsky & Shostakovich 10 Reviews A gentle giant who stops time, and an incessant steamroller of a Scherzo that gave birth to silence.
- 19 Oct 2026 Do a Model’s Conclusions Follow Its Reasons? Investigations In a pilot, Claude Sonnet 5.5 kept its conclusion on a public-safety case when the reason it verbalized was defeated, and switched to a different kind of reason.
- 10 Oct 2026 KPHP Tangents A classical radio station with one listener, run on WHRB’s afternoon-concert rules.
- 23 Sep 2026 Film emulation with f2palette Tangents Making raw photographs look as if they were shot on film, without letting a model touch the final pixels.
- 01 Sep 2026 PrairieGrader Tangents A grader for essay exams in which the page I grade on is the page the student gets back.
- 27 Oct 2025 Paraphon Tangents Every performance at the 2025 Chopin Competition, aligned so that one can switch pianists mid-phrase.