Aligning Values: Why AI Labs Are Hiring Philosophers and Historians
Understand why AI laboratories hire philosophers and historians to translate human values into mathematical constraints, protecting enterprise models from critical alignment failures.
Brewed for Work | Issue #4, June ‘26 | Premium
Welcome to Brewed for Work, your Tuesday ritual for sharper thinking about careers, workplaces, and the shifting nature of professional life.
If you’re reading this as a free subscriber, consider upgrading to the paid tier for full access to deep-dive essays, exclusive workbooks, and interactive polls. Your subscription keeps this independent newsletter fully brewed and your career better caffeinated.
» » » Check out subscribers’ benefits.
In this issue of Brewed for Work, we examine a fundamental shift in how artificial intelligence systems are built and deployed.
Technical founders frequently treat model safety as a standard debugging problem, assuming all solutions lie entirely within data pipelines. However, contemporary development has reached a boundary where purely statistical adjustments fail to resolve complex alignment issues.
We explore exactly why leading laboratories are currently hiring philosophers and historians to serve as ethics engineers. These professionals directly translate qualitative human values into rigorous algorithmic constraints.
Understanding this interdisciplinary approach provides a distinct competitive advantage for businesses scaling advanced software today.
Today’s Issue at a Glance:
The Shifting Skill Profile in Artificial Intelligence
Defining the Ethics Engineer
The Advantage of a Humanities Background
Transitioning from the Humanities to AI Safety
The Sustained Need for Humanistic Rigour in AI
When building artificial intelligence systems, technical founders frequently treat safety as a standard debugging problem. They assume that if a model produces biased or dangerous outputs, the solution lies entirely within the data pipeline or the loss function parameters. However, contemporary artificial intelligence development has reached a boundary where purely statistical adjustments fail to resolve deep behavioural alignment issues. This reality has forced leading artificial intelligence laboratories to change their hiring strategies, bringing philosophers and historians into engineering teams to serve as ethics engineers.
For leaders navigating the deployment of large language models, this shift is practical rather than philosophical. When you optimise a machine learning model using Reinforcement Learning from Human Feedback, you are fundamentally translating human values into a mathematical objective function. If the optimisation targets are poorly defined, the system will exploit loopholes in the reward architecture, a phenomenon known as Goodhart’s Law. Philosophers excel at identifying these logical vulnerabilities because their training focuses on the structural consistency of ethical frameworks.
Similarly, historians provide critical value by analysing patterns of human data and institutional behaviour. Large models are trained on massive historical datasets, meaning they inherently ingest the structural biases, past systemic inequalities, and linguistic shifts of previous decades. A historian understands how to audit these training corpuses, contextualising the information and anticipating how historical precedents will manifest in model outputs.
For an organisation, hiring humanities professionals is not a public relations exercise. It is a rigorous operational strategy designed to prevent catastrophic model failure, mitigate regulatory risks, and construct robust optimisation constraints that engineers can translate into code.
Integrating non-traditional specialists into technical workflows provides a distinct competitive advantage for businesses scaling artificial intelligence applications. By bridging the gap between qualitative human values and quantitative algorithmic constraints, these professionals protect enterprise systems from alignment drift.
