A philosophy PhD used to mean a long shot at a one-year teaching post and a drawer full of rejection letters. It can also bring a call from an AI company, offering six-figure pay to help decide how a chatbot should respond when someone tells it something painful. Firms once staffed almost entirely by engineers are hiring people trained to argue about right and wrong, and putting them in charge of how the product answers.
At Anthropic, that person is Amanda Askell, the philosopher who shapes how its chatbot, Claude, behaves. Beyond helping Claude handle emotional conversations with more care, she helps write the document that lays out how it should tell right from wrong. That document reads like a code of conduct for a machine, meant to guide millions of conversations every week. The Wall Street Journal reported that her job is “to teach Claude how to be good.”
Wired has reported a far more skeptical view from inside academia, where philosophers who take these jobs meet real suspicion. The suspicion comes down to timing. By the time a philosopher joins, the biggest decisions are usually made: what the product should do, who it should please and how fast it should ship.
The doubt goes deeper than timing. Philosophy earned its authority by having no employer. A philosopher answerable to nobody can tell a government or a company it has done wrong without risking a job. Put that person on a payroll and the freedom narrows, whether or not anyone asks them to change a single word.
Critics have a name for asking hard questions loudly enough for the public to notice, while making sure the answers never cost the business anything. They call it ethics washing. For a company under public pressure over how its AI behaves, hiring a philosopher can end up doing that job, whatever anyone intends.
Even a philosopher given real freedom, Wired reports, works for people who answer to investors. If an honest answer would slow a launch or cut into revenue, the company’s duty runs to its funders. The philosopher can raise the concern. The decision belongs to people the philosopher does not manage and cannot outvote.
Both readings of the arrangement, genuine safeguard or convenient cover, are defensible. The industry has not settled whether a flawed philosopher on the payroll still beats no philosopher in the room at all.
