“The end of learning is to repair the ruins of our first parents,” wrote John Milton in 1644. The image is difficult to repair: education as repair, as recovery, as restoration of capacities diminished by sin and neglect.
Four centuries later, in the age of Generative Artificial Intelligence (AI), that image has become essential again – as we are now surrounded by a technology that offers, on demand, the work we had long assumed we would have to do ourselves for education.
I came across Milton’s passage by chance when I was browsing a collection of the English author’s works and opened it to his 1644 tract. of education. Milton was not writing about algorithms. Yet he saw with unusual clarity the educational error that AI is now exacerbating: the confusion of language with learning.
Language, he wrote, “is but a tool that tells us useful things to know.” He warned against mistaking the command of words as a possession of the concrete things they are meant to represent. He linked language with substance, sequence with maturity, and study with direct contact with reality – principles which four centuries have made no less essential.
No technology in recent memory has made the device this large. Large language models like ChatGPT can summarize books, draft essays, organize research notes, translate paragraphs, generate code, and copy the prose that schools and universities have long taken as evidence of education.
They can be really useful when used with discipline. A professor can use them to prepare discussion questions. A researcher can use them to survey the literature more quickly. An administrator can use them to speed up routine writes. It would be foolish to deny their usefulness.
But utility is not the same as learning, and AI amplifies an old weakness. This leads us to mistake verbal fluency for the same. A student can produce sophisticated prose without really struggling with the question. A researcher can prepare a competent summary without seeing the problem clearly. A professional may seem informed without making any decisions. The danger is not just of dishonesty – it is of substitution.
For Catholic education, that replacement matters because learning is not the production of acceptable performance but the formation of a person capable of truth, judgment, and responsibility.
Milton saw a version of this in his own time. He criticized the practice of demanding “contents, verses, and speeches” from young students even before their minds had been formed by “long reading and observation.” He objected to demanding ready performance before the underlying powers were mature.
Generative AI industrializes exactly the same academic mistake. It provides the student with ready-made language before it goes through the reading, questioning, hesitation, and revision that makes the language meaningful. What Milton considered the error of sequence, AI turns into a system.
This matters because education is not made up of answers alone. Every teachable answer was once an answer to a question someone actually asked.
Students do not assimilate knowledge simply by drawing conclusions – they must be brought into question. That is why the main agent of education is the student. No one can learn from someone else. A device can assist in instruction; It cannot do any educational work for the student.
The role of the teacher in the age of AI accordingly becomes more important, not less. A real teacher is not just a distributor of content. A real teacher is an experienced guide in questioning: someone who knows what the student has not yet seen, what distinctions need to be made, what confusion needs to be exposed, and what question should come next. Best class is not a transfer of information from one container to another. It is a living act of thought. This is why seminars, debates, laboratories, tutorials and serious conversations retain their influence even when the information itself becomes cheap.
We celebrate knowledge: facts gathered, results confirmed, information stored. But as biologist Stuart Firestein has argued, discovery begins not only with what we know, but also with a disciplined sense of what we do not yet understand. That limit is where large language models reach their limits. They can reorganize collections with amazing fluency, but they cannot live in uncertainty, raise a genuinely new question, or take responsibility for the truth.
This explains why some tasks cannot be assigned to machines without preventing them from occurring. Paying careful attention to a text, weighing conflicting evidence, deciding whether a conclusion is warranted or not, taking responsibility for what one claims – these are not helpful tasks. These are the tasks that create the mind.
No machine can perform them in our country – not because machines lack processing power, but because these tasks have no impact unless a person performs them. Their purpose is not to generate output. It’s up to the creator to make them do them.
Education worthy of the name has always understood this. Its end is not the delivery of content, no matter how accurate. It is the creation of individuals capable of judgment, focus and intellectual honesty. That formation requires a real encounter with difficulty – the friction of a difficult text, the resistance to a problem that does not quickly yield results, the discomfort of revising what one believes. It requires intellect as much as embodiment: reading slowly, speaking in your own voice, accepting the cost of standing behind your words. Mere management of information does not make a person capable of truth. Intelligence is created in contact with reality, not in imitation of it.
So the deepest challenge of AI in education is not academic integrity, although that problem is real. The question is whether we will allow our schools and universities to define learning in terms of the production of acceptable outputs. If this is our standard, outsourcing will always look like efficiency. But if education is a creation of judgment, then replacement becomes self-defeating.
What should institutions do? The answer is neither panic nor complete prohibition. This is a new educational format. More writing was done in class. More verbal defense of arguments. More seminars were organized around live questions rather than passive downloads of information. More laboratory and studio work that requires students to explain not only what the result shows but also what it does not show.
When students use AI, a reasonable requirement is transparency: disclose what was asked, what the system produced, what was kept, what was rejected and why. The issue is not about surveillance. This is intellectual ownership – the habit of standing behind your thinking. Institutions must also reinvest in teacher-scholars whose presence, judgment, and intellectual rigor cannot be automated.
The same commitment is there at home also. Device-free dinner tables, conversations across generations, reading out loud together, and the habit of asking kids not only what they think but also why – these are little schools of freedom. They teach that education is not the production of impressive sentences. It is the creation of honest minds.
In the light of the moment we are living through, there is less crisis than explanation. AI has not created new educational problems; This has made the old impossible to ignore. The habit of rewarding performance over understanding, fluency over depth, and polishing over genuine engagement was already present in our institutions before the first language models were trained. AI simply industrializes and accelerates those habits until their emptiness becomes undeniable.
That may be its most unexpected gift. If this disruption forces us to reclaim what education was always meant to be – the formation of minds capable of asking genuine questions, making careful decisions, and taking responsibility for the truth – then the age of AI may, paradoxically, prove to be an age of educational renewal.
Milton’s deeper claim goes further. The end of learning is not mere competence or civic virtue, but “knowing God rightly, loving Him, imitating Him, becoming like Him”. In that sense, education participates in the restoration of what sin has obscured.
No machine will be able to repair those ruins. That restoration is ultimately God’s work before it is ours; Yet, with the help of grace, we must still perform the human labor of attention, judgment, and love.
Santiago Schnell is Provost and Professor of Mathematics at Dartmouth, and holds adjunct appointments in Biochemistry and Cell Biology and Biomedical Data Science at the Geisel School of Medicine. A mathematical biologist by training, he also writes on the Catholic intellectual tradition, the philosophy of science, and the mission of Catholic higher education.
<a href