The Medieval Mind in the Age of AI

Why the most advanced technology we have built may require an older way of thinking

Modern people tend to imagine history as a slow escape from ignorance. Medieval life, viewed through contemporary eyes, can seem dominated by superstition, mystery, folklore, and a lack of scientific understanding. We call entire periods “dark ages,” implying that humanity simply knew less.

Yet this picture misses something important.

For centuries, people lived in worlds they could they could skillfully navigate but could not fully explain. A medieval farmer understood the rhythms of the land, when to plant crops, how weather patterns affected harvests, and how to survive difficult winters. A blacksmith could forge tools of astonishing durability without understanding metallurgy or molecular structure. People possessed deep practical knowledge of their world, even when the mechanisms behind it remained mysterious.

Modern society changed this relationship. Through science and technology, we developed a new faith: not that we understand everything, but that everything is, in principle, understandable. Even when we do not know how something works, we assume someone does. As the sociologist Max Weber argued, modernity replaced mystery with the belief that the world could ultimately be mastered through explanation.

Artificial intelligence may challenge this assumption.

Systems like ChatGPT and Claude are, in many ways, the culmination of scientific rationalism: products of mathematics, engineering, computation, and unprecedented scale. Yet they present us with a strange paradox. While we understand how to build and train these systems, we still struggle to explain how many of their capabilities emerge. They are tools we can use remarkably well without fully understanding why they work.

Perhaps this is the first technology of the modern age that weakens our confidence that explanation will always catch up.

If so, the age of AI may require something we thought we had left behind: a more medieval relationship to knowledge - one grounded less in certainty and more in learning how to live with mystery.

Living With Mystery

To understand our relationship with AI, we first need to look to the past.

Modern people often imagine pre-Enlightenment societies as worlds of mystery, folklore, and limited scientific explanation. Medieval life, viewed through contemporary eyes, can seem irrational and fearful, populated by people who simply lacked the scientific tools to understand the world around them. We picture societies trapped in darkness, waiting for reason to finally arrive.

But this view misses something important.

The medieval person did indeed lack modern scientific explanations. A farmer could not describe photosynthesis, soil chemistry, or microbial ecosystems. Yet they understood planting seasons, weather patterns, crop rotation, and the practical rhythms of survival with remarkable precision. A brewer could master fermentation without knowing yeast existed. A blacksmith could forge durable tools and weapons without understanding metallurgy.

This was not ignorance so much as a different relationship to knowledge.

Medieval people often possessed deep practical understanding without abstract explanation. Their knowledge was local, embodied, and inherited through lived experience rather than scientific models. They knew how to work with the world, even when they could not explain its underlying mechanisms.

Just as importantly, mystery itself was not considered a problem to be solved.

Many aspects of reality remained beyond explanation: storms, illness, fate, and the wider structure of the cosmos. Yet these unknowns were not always experienced as failures of understanding. Medieval people lived in a world where mystery was expected rather than treated as something that demanded explanation. The world was often experienced as enchanted, alive with forces, meanings, and spiritual significance that modern society tends to separate from everyday life.

This does not mean medieval life was idyllic, rationally superior, or free from suffering. Few of us would willingly trade modern medicine or scientific progress for life in the Middle Ages. But we may misunderstand what it meant to live without scientific explanation. A lack of abstract knowledge did not necessarily produce fear. It could also produce familiarity, practical competence, and a greater comfort with uncertainty.

Medieval people understood far less about the universe as a whole. Yet they may have been more comfortable living alongside mystery. Modernity would eventually change this relationship by replacing mystery with something new: the belief that, in principle, everything could be explained.

Everything is Explainable

Modern life is full of technological miracles that barely register as miraculous. You are likely reading this on a smartphone or laptop built through fabrication processes so complex that no single individual fully understands them. Satellites orbiting the Earth help determine your location within meters. Airplanes carry us across continents with such reliability that we become irritated when the Wi-Fi stops working.

What is striking is not simply that these technologies exist, but how quickly we stop experiencing them with wonder. We do not treat them as mysterious. We expect them to work.

The sociologist Max Weber believed that this shift reflected something deeper than technological progress. Modernity, he argued, changed not only what we knew, but our relationship to what could be known. Scientific progress gradually replaced mystery with what Weber called disenchantment:

“The knowledge or belief that if one but wished one could learn it at any time… there are no mysterious incalculable forces that come into play, but rather that one can, in principle, master all things by calculation.”

The significance of this claim is easy to miss. Weber was not arguing that modern people personally understand the world better than those who came before them. Most of us could not explain electricity, semiconductors, financial systems, or how information travels across the internet. In fact, psychologists have repeatedly shown that people are often remarkably poor at explaining even simple everyday objects such as toilets, zippers, or bicycles despite feeling confident that they understand them.

Modernity did not eliminate ignorance. It changed how we think about it.

In medieval life, mystery was often accepted as part of reality. In modern society, mystery became temporary ignorance, a problem waiting for explanation. Even if I do not understand something, I assume that someone, somewhere, does.

This shift was only possible because modern society increasingly relied on abstraction, standardization, and scale.

Abstraction allows us to use tools without understanding how they work. I do not need to be a mechanic to drive a car, an engineer to use a smartphone, or a pilot to board an airplane. Knowledge is delegated to specialists, institutions, and systems. This is not a weakness of modernity but one of its greatest strengths. If everyone needed to master every tool they used, progress would grind to a halt.

But abstraction only works because the modern world is highly standardized and scaled. Cars, aircraft, medicines, electrical systems, and communication networks function because agreed protocols and standards make them predictable and reliable. I can board a plane without understanding aerodynamics because I trust the broader system of engineering, regulation, and expertise behind it.

Scientific rationalism succeeded so completely that we stopped noticing its assumptions. We became comfortable relying on tools we could not personally explain because we trusted that explanation existed somewhere. Mystery had not disappeared, but it had been transformed into a temporary problem awaiting explanation.

And so we stopped relating to technology with wonder. We expected it to work.

But this confidence depended on something subtle: the belief that, if we truly wished, someone could ultimately explain how everything around us worked in a clear, rational and scientific manner.

Yet the culmination of this scientific worldview may also be the first thing to destabilize it. AI is not simply another tool. It may be the first tool that forces us to confront the possibility that explanation itself has limits.

The Return of Medieval Craft

As we have already seen, modernity made us comfortable relying on tools we do not personally understand. We use smartphones, airplanes, financial systems, and electrical grids without hesitation because we assume that explanation exists somewhere. Even if I cannot explain how something works, someone, somewhere, can.

Artificial intelligence feels different.

A jet engine is enormously complicated. A financial market is highly complex. The distinction matters. Complicated systems may contain countless interacting parts, yet we still feel their mechanisms ultimately “bottom out” in engineering and physics. Complex systems, like markets or weather, may resist prediction because of their scale and interdependence, but we do not usually regard them as fundamentally mysterious. Their unpredictability does not imply a failure of explanation.

AI seems to occupy a different category.

Models like ChatGPT and Claude are, in many ways, the culmination of scientific rationalism itself. They are built from mathematics, engineering, computation, and unprecedented scale. Trained on astonishing amounts of human data using immense computational resources, they represent the triumph of abstraction and scientific progress.

Yet paradoxically, they also challenge one of modernity’s deepest assumptions.

Previous technologies abstracted physical labour, transportation, memory, or calculation. LLMs increasingly abstract something far more intimate: thinking itself. The very cognitive capacities we ordinarily use to understand tools, reasoning, language, conceptual understanding, are now partially outsourced to the tool itself. In trying to explain AI, we increasingly find ourselves confronting questions about thinking that we do not fully understand even in ourselves.

While researchers increasingly understand how to build and improve these systems, many aspects of their behaviour remain impossible to explain. Capabilities emerge unexpectedly. Internal representations remain opaque. Researchers still debate the extent to which models form world models, reason abstractly, or develop internal concepts. What makes AI feel different is not merely its complexity, but the possibility that some aspects of its behaviour may resist the kind of clear, rational explanation modernity taught us to expect.

This is not an argument against science, interpretability, or rigorous inquiry. Far from it. If anything, systems this powerful demand even greater scrutiny. We should continue probing these models, testing them, improving them, and studying their behaviour. Accepting limits to explanation is not the same as abandoning rigor.

But perhaps the mindset we bring to these systems needs to change.

Modernity taught us to relate to technology through explainability. If something is built, then surely it can ultimately be understood, calculated, and rationally described. The medieval mindset offers another possibility: practical engagement without complete explanation.

A craftsperson rarely relates to their work through explicit rules alone. A carpenter develops feel for wood grain through repeated engagement. A pilot may understand the physics of flight, but ultimately learns to fly through hours in the cockpit. Skill emerges through a mixture of explanation, intuition, experimentation, and experience. The goal is not perfect quantification, but practical fluency.

In some ways, we are already beginning to work with AI like this. We prompt models, test boundaries, compare outputs, observe failure modes, and gradually develop intuitions about what works. Researchers experiment with behaviours, steering techniques, and interventions to better understand how models respond in different contexts. This is not blind faith or mysticism. It is disciplined experimentation: learning through interaction with something only partly understood.

Perhaps the challenge of AI is not to abandon explanation, but to loosen our expectation of total explainability. We may need to become less like architects of total understanding and more like craftspeople learning to work skillfully, critically, and responsibly alongside systems we only partly understand.

The age of AI does not require a return to superstition or magic, but perhaps to something older and more human: the ability to live responsibly alongside mystery.

The Return of Meaningful Engagement

The age of AI does not require a return to superstition or magic, nor should it. We should continue to question, test, interpret, and improve these systems with all the scientific tools we used to get this far. But perhaps AI is asking something unexpected of us, not to abandon explanation altogether, but to become more comfortable with its limits.

Medieval people understood far less about the broader universe than we do, yet they often lived in closer contact with the immediate world around them. Their relationship to knowledge was practical, situated, and deeply tied to the rhythms of everyday life. They learned through engagement, repetition, and feel.

Modernity gave us extraordinary explanatory power, but often at the cost of intimacy with the systems we depend upon. We increasingly inhabit abstractions in the form of institutions, technologies, and invisible infrastructures that work remarkably well but force us to be distant from experience.

Perhaps the strange irony of AI is that one of the most advanced technologies we have ever built may gently pull us back toward something older: a more engaged relationship with the our everyday world. Less obsessed with total explanation, more attentive to practical fluency. Less focused on mastering every mechanism, more willing to learn through participation.

The endpoint of the Enlightenment may not be total understanding, but a renewed encounter with something modernity taught us to distrust: mystery. Not mystery as ignorance, but mystery as a reminder that explanation is not the only path to understanding. Some forms of knowledge emerge through participation, practice, and feel, through learning by doing. In learning to work skillfully alongside what we cannot fully explain, we may rediscover not only a more meaningful way of engaging with AI, but a more immediate relationship with the world around us.