Philosophy on the payroll

The most powerful technology companies on the planet are hiring philosophers at a steady clip, and the statistic cited by The Economist is itself a small intellectual scandal: today, in the United States, a philosophy graduate is more likely to find work than a computer-science graduate. The discipline that for decades was the favourite target of people who liked to humiliate humanities students has become Silicon Valley’s most sought-after life insurance.

The reason is easy to state and hard to accept: artificial intelligence poses, in the magazine’s own words, “thorny problems, the kind philosophers prefer.” What is consciousness? Can a language model suffer a moral wrong? What do we owe an entity we cannot tell whether it feels anything? These are questions an engineer is not trained to ask, and that a public company — or one about to go public — cannot afford to leave unanswered, not because it fears being ethically wrong, but because it fears having no answer ready when someone asks in a hearing, in court, or in a newspaper.

Are these philosophers still philosophers, or have they become subtle mediators between academic credibility and business necessity? Amanda Askell, with a New York doctorate on ethics of infinity, now leads personality alignment for a chatbot. Joe Carlsmith, Oxford doctorate, worked on another model’s “constitution” — an eighty-four-page document that cites Kant, the Universal Declaration of Human Rights, and Apple’s terms of service together, perhaps the most involuntarily comic synthesis of our time: the categorical imperative beside the privacy agreement nobody reads. Henry Shevlin and Atoosa Kasirzadeh work on DeepMind’s “readiness” for AGI.

Sam Altman has claimed to have consulted “hundreds of moral philosophers” to write ChatGPT’s behavioural rules — a claim no journalist has ever managed to verify by simply asking for the names.

At this point one has to ask whether that critique has ever had consequences: whether it has ever cost a delayed launch, a removed feature, a changed product. A company can tolerate internal critics without granting them any real power. Authorised dissent is often the most sophisticated form of controlling dissent, because it lets the company say “see, we have people who contradict us” while that voice remains structurally without leverage. The question to put to every corporate philosopher is not “have they ever silenced you?” but “have you ever won?” — and if the answer is “I don’t know, it’s confidential,” the suspicion is already an answer.

There is a second, more uncomfortable level. Edward Harcourt, a philosopher at Oxford, has spoken openly about the risk of ethics-washing: ethics as varnish, not architecture. Those who do this work inside companies would reply — and with some reason — that the charge of “decoration” is too convenient, because it never recognises the case where philosophical work genuinely counted. Perhaps that is true. The problem is that, from outside, we cannot tell the two cases apart: the opacity of corporate decisions makes the question empirically unresolvable, and that unresolvability is already part of the problem. A discipline that asks the world for transparency should be able to show its own.

There is a historical irony worth remembering. For centuries medieval philosophy was called ancilla theologiae, handmaid of theology: free to reason, but only up to the point where reason confirmed the faith. Today it risks becoming the handmaid of capital: free to reason, but only up to the point where reason confirms the business model. The difference is that no one in the thirteenth century pretended theology was neutral. Technology companies, by contrast, insist on presenting ethics as pure research while paying its salary.

This does not mean these philosophers are acting in bad faith, nor that their work is useless — that would be as lazy an accusation as the opposite apology. It means only that the value of their work is not measured by their individual good faith, which is almost always genuine, but by a structural fact no personal biography can correct: whoever pays the salary remains, in the end, the audience that must be answered first.