Singer at eighty, with an AI of his own: has the good-life question changed?

A philosopher making a copy of himself is a new event in the history of philosophy.

On 30 October 2024 Peter Singer announced on his social media accounts that a chatbot trained on his own writings — Peter Singer AI — had gone live.

The system was developed by Sankalpa Ghose in dialogue with Singer. His writings, his talks and the features of his manner were placed, by way of detailed prompt engineering, inside a retrieval-augmented generation architecture. The aim was to produce an interlocutor that questions the user in the Socratic manner, asking something back rather than handing over an answer.

The Cambridge Quarterly of Healthcare Ethics has published an interview with the machine as an academic article: “Sentience and Beyond — A Representative Interview With Peter Singer AI.”

A philosopher who has spent fifty years arguing for the widening of the moral circle has now extended himself into a piece of software. The intersection is worth pausing over.

Back to Melbourne#

Singer was born in Melbourne on 6 July 1946; this summer he turned eighty. After twenty years at Princeton he has returned to Australia.

The Guardian's interview of 14 August 2026 finds him walking along Merri Creek, north of Melbourne. He is not retired: recent years have brought a podcast, a global alliance for corporate philanthropy and a version of himself in artificial intelligence.

The headline of the interview is of the sort that will start an argument on the left:

“Capitalism is still the most productive economic system that ever existed.”

Where he stands on capitalism#

The sentence does not mean Singer has fallen into inconsistency. On the contrary, it makes plain a tension effective altruism has carried from the beginning.

His proposal is not to abolish the system but to direct a substantial part of the economic power it generates towards solving social problems. The current instance of this is Profit for Good, of which he is a co-founder: a growing global alliance of companies that undertake to give at least ten per cent of their net profits to high-impact charities.

The position attracts serious criticism, and it should.

The objection from the left: reducing poverty to a matter of donation means leaving the structure that produces it unquestioned. A company that gives away a tenth of its profits may thereby legitimate how it made them.

Singer's reply is predictable and consistent. Changing the structure takes a long time, and in the meantime preventable deaths go on. To sacrifice a life that could be saved today to a revolution due tomorrow is not a moral choice.

The argument is not settled. But it is clear that both sides have a serious case.

The real question: how should we live?#

Singer's books — Animal Liberation, Practical Ethics, How Are We to Live?, The Most Good You Can Do — make up a body of work that takes ethics out of academic specialism and ties it to everyday decisions.

But Singer's real break lies not in any thesis; it lies in the way he frames the question.

In the modern period, philosophy has mostly asked how we should live by way of our own lives: happiness, autonomy, self-cultivation.

Singer translates the question as follows: is a good life only a life that is good for us?

The consequences of that translation are uncomfortable, and Singer makes them so deliberately. If you accept that you must ruin an expensive pair of shoes to stop a child from drowning, what grounds have you for withholding a donation that would prevent the death of a child far away? Does distance make a moral difference?

Singer's power lies not in his answer but in his making the question unavoidable.

Wanting to help is not enough#

The second distinctive move in Singer's ethics is his insistence that good intentions are not sufficient.

With the resources we have, we ought to try to bring about the greatest good possible. That means a philanthropy grounded in evidence and reasoning: measuring how many years of life a given intervention saves per pound given.

The approach became one of the intellectual sources of the effective altruism movement — and the crises that movement has suffered in recent years have fed part of the criticism aimed at Singer.

But the principle of measurement itself remains defensible. When you give money to a charity, asking what that money does is a reasonable demand.

Why artificial intelligence is a special matter for Singer#

At the centre of Singer's work stands the concept of sentience. For a being to merit moral consideration it need not be intelligent; it is enough that it can suffer. That criterion, taken from Bentham, is the contemporary foundation of animal ethics.

It applies directly to artificial intelligence. Which is precisely why Singer occupies an unexpected position within the debate about consciousness that we covered today: the logic of widening the moral circle can be run so as one day to take in machines as well.

That Singer has built his own artificial intelligence sharpens the question further. Peter Singer AI is a text engine generated from Singer's writings — not a person. But its design as an interlocutor that answers a question with a question, in the Socratic manner, presses on the limits of the notion of an interlocutor.

A philosophy that gives no comfort#

Singer's philosophy does not comfort the reader, because it takes morality out of abstract principle and attaches it to everyday choice.

What do we buy? How much do we consume? Where does our money go? To what extent do we count the suffering of others in our own reckoning?

The difference between Singer and Krista Lawlor's analysis of reasonableness, which this column covered in recent weeks, lies exactly here. Lawlor proposed that we understand why people think differently. Singer reminds us that thinking differently has a price.

Both are necessary. The second is the less popular.

---

In Turkish: many of Singer's books have been translated, among them Animal Liberation, Practical Ethics, One World, The Most Good You Can Do and Applied Ethics.