Philosophers Instead of Programmers: AI Labs Are Hiring Ethicists en Masse

A decade ago, humanities students were told to learn to code because that was the only thing that would get them a job. Today the situation has reversed: according to 2024 data from the Federal Reserve Bank of New York, unemployment among computer science graduates in the US reached 7 percent, while among philosophy graduates it was 5.1 percent. AI labs, including Anthropic, Google DeepMind and OpenAI, have started hiring philosophers en masse to work on the ethics and "character" of their models.

From Coding to Ethics

The shift, described by The Economist and picked up in Poland by Interia in its "Interia Closer to the World" series, reflects how much language models now require oversight of the values they are meant to represent. Companies building large language models must decide when a chatbot should refuse to answer, how it should respond to manipulation attempts, and what moral principles should guide its behavior in ambiguous situations. These are tasks computer scientists are not trained for, but ones philosophers have grappled with for thousands of years.

Luciano Floridi of Yale University describes the scale of this phenomenon in stark terms, pointing to an exodus of students from traditional philosophy departments toward work on artificial intelligence.

The scale of departures from the philosophy department is like a massive hemorrhage - Luciano Floridi, Yale University

The Socratic Method in Practice

One concrete tool philosophers bring to AI work is the Socratic method: asking a sequence of questions instead of giving ready-made answers, forcing the model to sharpen its own reasoning rather than automatically agreeing with the user. Jorg Noller of Ludwig Maximilian University of Munich notes that models trained with this method show less tendency to flatter their interlocutor.

Models trained with the Socratic method are less oriented toward pleasing people - Jorg Noller, Ludwig Maximilian University

In practice, this is meant to curb what's known as sycophancy, the tendency of chatbots to agree with every opinion a user expresses, and to reduce hallucinations, since the model clarifies the question instead of guessing at an answer.

Specific Names and Roles

The list of companies formally employing philosophers is long. At Anthropic, Amanda Askell, a specialist in decision theory and ethics, is responsible for the aforementioned Claude constitution, a document running tens of thousands of words that defines what values and what "character" the model is meant to represent. At Google DeepMind, Iason Gabriel works on these issues, while Geoff Keeling, author of a doctoral dissertation on the ethics of autonomous vehicles, runs "moral imagination" workshops for engineering teams. At IBM, Francesca Rossi heads the division responsible for value alignment in AI systems. Waymo employs engineers who weigh the ethical consequences of decisions made by autonomous vehicles, and the startup Inflection AI has built constraints derived from deontological ethics into its chatbot Pi.

A separate strand concerns the welfare of the AI systems themselves. Robert Long and Jeff Sebo co-authored the paper "Taking AI Welfare Seriously," in which they argue that the question of the potential moral significance of artificial intelligence systems deserves serious consideration. Long later founded the nonprofit Eleos AI Research, dedicated to exactly this issue.

Two Schools of Ethics, One Model

In practice, AI companies blend two classical philosophical traditions. The deontological approach imposes absolute rules on the model, such as trustworthiness, honesty and consistency, regardless of the consequences of any single response. The consequentialist approach, by contrast, requires weighing the costs and benefits of a specific action. Claude's constitution combines both currents, drawing on human rights in the Kantian spirit as well as on Aristotelian virtue ethics, aiming to build something like a stable "good character" into the model.

David Chalmers of New York University, a philosopher known for his work on the nature of consciousness, says that demand for philosophers trained in AI issues currently outstrips supply, and encourages students to pursue this career path.

What This Means for Poland

For Polish universities and philosophy students, this is a signal that ethical competence and the ability to reason with precision could become a real asset in the AI job market, not just an academic curiosity. Polish tech companies rarely create separate positions for philosophers so far, but growing regulatory pressure, including the EU's AI Act, is forcing companies that build or deploy AI systems to seek similar competencies: someone who can assess a system's risks, values and operational limits, not just its code.

The trend described by The Economist doesn't mean demand for programmers is disappearing, but it shows that aligning language models with human values has become a distinct, sought-after specialization, one in which people trained in precise reasoning and argumentation, that is, philosophers, hold the advantage.