Philosopher · Researcher · Writer
PhilosophyAIMindEthicsTechnology
Move through the field. Each concept is a live question: hover one to see what it touches, open it to follow the thread.
I am a philosopher working on knowledge, testimony, consent and responsibility, and on what becomes of each when the source or the decision-maker is a machine.
My area of specialisation is epistemology, with a focus on testimony and the social epistemology of artificial intelligence. I argue that large language models, however reliable, do not on the best current evidence transmit testimonial knowledge: what they lack is not a mind but a connection, running through their own operation, between the truth of what they say and why they say it. Their users stand to them as readers of an instrument, not as hearers of a speaker.
I also work on the ethics and politics of algorithmic systems. Taking India's Aadhaar as the paradigm case, I argue that where an algorithmic system produces the standing of the people it governs, their consent cannot legitimate it; its legitimacy rests instead on obligations of legibility, non-abandonment and remediation that no one can waive. I am an independent researcher in New Delhi, and studied philosophy at the University of Delhi.
Computational systems built to perform tasks we would call intelligent if a person performed them, such as answering questions, translating, classifying and planning.
AI challenges Mind: AI is the first serious empirical pressure on our concept of mind since the discovery of the brain.
Whatever it is that perceives, thinks, feels and acts; the philosophy of mind asks what that is, how it relates to the body, and what else might have one.
Mind grounds Agency: Minds are the paradigm bearers of agency; the question is whether they are the only ones.
The capacity to act: to do something, rather than merely have something happen, for reasons that are in some sense one's own.
Agency requires Knowledge: To act for a reason you must take something to be the case; agency presupposes at least belief-like states.
True belief held in a way that makes its truth no accident; what exactly secures that has been disputed since Gettier.
Knowledge grounds Ethics: What we owe each other epistemically (honesty, care, the sharing of evidence) is where knowledge becomes ethical.
The study of how we ought to act and what we owe one another, including what institutions owe the people their systems govern.
Ethics grounds Responsibility: Responsibility is where ethics becomes practical: who answers for what was done.
The relation between an agent and an outcome in virtue of which the agent can rightly be praised, blamed or asked to answer for it.
Responsibility challenges Technology: Systems that learn and act at scale distribute causal contribution so widely that traditional responsibility seems to find no one to attach to.
The made world of tools, systems and infrastructures, considered not as neutral instruments but as arrangements that shape what is possible, and for whom.
Open problems, some of them worked out in papers. Each one has a short answer here and a full one on its own page, and leads into a research area, a paper, an essay or an experiment in the lab. All questions
Not in the sense that matters for knowledge. On the best current evidence, large language models do not transmit testimonial knowledge: the truth of what they say does not figure, through their own operation, in why they say it. Their answers are often true and often useful, but relying on them is closer to reading an instrument than to taking someone's word.
Where an institution produces its subjects' operative standing as entitlement-bearers, its legitimacy depends not on their authorisation but on discharging three obligations they cannot waive: legibility (recognition criteria that fit the material reality of their lives), non-abandonment (not withdrawing the standing it produced without meeting a burden of justification), and remediation (a real path, inside the institution, when recognition fails).
On the best current evidence they probably do not, though the question is open in a precise and testable way. Their internal states carry structure that correlates with truth, and that structure can be used to make outputs more truthful; what has not been shown is that it governs what they assert when nobody intervenes. Knowing requires the second.
Responsibility falls mostly on you and on the people who built and deployed the machine, though not in the way a speaker is responsible for what she tells you. When you rely on a language model, the warrant for your belief is generated by your own use, so the burden of checking falls on you. Its makers answer for putting it into the world and for how it behaves, but responsibility for a system is not authorship of its sentences.
A system may act for reasons without being conscious; the question is open. The features that make behaviour count as action can each be specified without mentioning experience: responsiveness to reasons, flexibility across circumstances and the capacity to revise. What has to be settled is whether a system that meets those specifications acts or only behaves as if it does, and I think it is best settled by asking what governs its behaviour rather than whether anyone is home.
Responsibility for harm done by an AI system moves; it does not disappear. The 'responsibility gap' worry is that no one controlled a learning system closely enough to be blamed, and the system cannot be. I think responsibility in these cases is redistributed to the people who design, deploy and oversee such systems, in the way we already hold the makers of instruments answerable, and the philosophical work is to draw that distribution accurately.
There may be no fact of the matter that settles it, and that may matter less than it seems. Philosophers long argued about teleporters and split brains as thought experiments; with software systems that are checkpointed, forked and merged, those cases have become routine operations. If Parfit was right that identity is not what matters in survival, the better question is which kinds of continuity we actually care about, and why.
Eleven areas that share questions. The lines are where one area's problem becomes another's.
Selected papers. The full archive can be searched and filtered by year, topic, venue and research area.
Episteme, 1–22
AI and Ethics 6, 414
AI & Society
Essays written to be read slowly: arguments developed at length, with footnotes, for readers who want more than a position.
Interactive instruments for classic and new problems. Set the parameters; see what the theories say; notice where your intuitions came from.
What are you actually measuring when you decide something has a mind?
Run experimentAt what point in a gradual replacement do you stop being you?
Run experimentWhere does responsibility for a belief live when it arrives through a machine?
Run experimentFor collaboration, review, talks, or disagreement, which is often the most useful.