AI and the End of Wisdom

Published in the European Conservative.

Anyone who belongs to a school or university will know that artificial intelligence is already having a major impact on education. Despite launching less than a year ago, swathes of students are now using ChatGPT to write their assignments and essays, with other AI tools also offering to transcribe notes and summarise the contents of readings. Even beyond the classroom, such technologies are rapidly revolutionising the nature of knowledge and how we learn. This is because, for the first time, they have made it possible to outsource mental processes—specifically, the remembering, processing, and reproducing of information—to inhuman entities. They are, to borrow the terms of the French philosopher and early critic of technology Bernard Stiegler, exteriorizing our mental faculties to non-living prostheses.

In an age where many transhumanist technologies are taken for granted—be it the birth control pill which circumvents women’s biology or modes of communication which transcend the natural limits of geography—many don’t see this as posing any existential significance. AI, they might argue, simply does menial or supposedly ‘pointless’ mental tasks on our behalf, and so allows us to get on with whatever we deem more worthwhile. Some even claim that AI allows us to be more creative for this reason. Yet, by dehumanising knowledge, it may ultimately inhibit the possibility of true wisdom. This is because wisdom is something that can only be realised through authentically human processes; at least, this is what we learn from pre-modern conceptions of knowledge, especially that of Plato.

In trying to understand the danger that modern technology poses to wisdom, we ought to revisit the Phaedrus, the dialogue in which Plato presents his theory of pedagogy and also his famous critique of writing—arguably the first technology to exteriorise knowledge to a non-living entity. In the dialogue, Socrates explains to his interlocutor that the art of rhetoric requires a teacher who can present ideas in such a way that aligns with the particular dispositions of each student; that is to say, he must be able to speak to their souls.

“He who is to be a rhetorician,” he says, must know the various forms of soul … men of a certain sort are easily persuaded by speeches of a certain sort for a certain reason to actions or beliefs of a certain sort, and men of another sort cannot be so persuaded. (Phaedrus, 271d)

The implication here is that the effective transmission of knowledge is necessarily an interpersonal activity: it requires living people to engage in a dialectic with other living people, working with the particularities of their nature and the degree of understanding that they already possess.

Having explained this didactic ideal, Socrates goes on to discuss writing as something which endangers it. He recounts the myth of its invention as follows: one day, Theuth—the ancient Egyptian deity said to have invented mathematics, geometry, and astronomy—went to Thamus, the King of Egypt, to present him with a new branch of technical knowledge that he believed would improve people’s wisdom and memory, namely the art of writing. But the king did not approve of his art, declaring that it would in fact have the opposite effect:

If men learn this, it will implant forgetfulness in their souls; they will cease to exercise memory because they rely on that which is written, calling things to remembrance no longer from within themselves, but by means of external remarks. What you have offered is a recipe not for memory, but for reminder. And it is not true wisdom you offer your disciples, but only its semblance. (Phaedrus, 275a)

Moreover, King Thamus says, writing is dangerous because it can “drift all over the place, getting into the hands not only of those who understand it, but equally those who have no business with it; it doesn’t know how to address the right people, and not address the wrong.” Considering what Socrates had said earlier in the dialogue regarding the imperative of the rhetorician to “know the various forms of soul,” we might interpret the concerns of King Thamus to mean that the danger of writing as a technology which removes interpersonal mediation lies precisely in the fact that it dehumanises learning. The concern is that writing might not only inhibit people’s ability to remember things—the importance of which we will turn to shortly—but it may also fail to address people by speaking to their particular soul. As such, it might undermine the very purpose of the Platonic endeavour, that is, to guide people to the truth within themselves through a process of remembrance which he called anamnesis.

For centuries, this pedagogical ideal was upheld by Christian, Jewish, and Islamic traditions of learning. Even when writing was adopted, it was still believed that its contents should be taught interpersonally, and in degrees corresponding to the capacities of each learner. Following the Enlightenment, however, these traditions were, so to speak, overwritten—and Plato’s warning was seemingly forgotten.

The Enlightenment’s emphasis on egalitarianism and universal reason led to the decline of particularised pedagogies, favouring instead a universalised, standardised approach to learning–one appropriate, so it was thought, to the new, emancipated, modern man. This change in our understanding of pedagogy culminated in the French Encyclopédist movement of the 18th century, an attempt to codify all knowledge into a singular body of information accessible to anyone, anywhere. In 1938, the English science fiction writer H.G. Wells foresaw the synthesis of encyclopedism with the emerging technologies of his time into what he called the “world brain”: a global index of knowledge independent of all human mediation and authority that would serve, in his own prophetic words, a “world culture” (the same world culture, perhaps, that Julian Huxley made the guiding philosophy of UNESCO just a few years later).

Within less than a century, Wells’ dream became a reality, thanks to an innovation that would surely have been the Egyptian king’s worst nightmare: the internet. In replacing interpersonal pedagogies with an inanimate body of information, the internet has exaggerated the dangers of writing to an unprecedented extent. It may appear to provide us with more knowledge than ever before, but—because it is disembodied and impersonal—it cannot speak to particular souls and so guide them to truth in a meaningful, deeper sense. To echo Plato’s concern, it does not offer true wisdom, but only its semblance.

It is the dynamic process of interpersonal dialectic that allows us to move from facts to rational principles, and then from rational principles to eternal truths; to move up the epistemic scale that is Plato’s “divided line.” This, essentially, is the danger of substituting technology for traditional modes of knowledge. Technology, being static and inhuman, cannot guide us to union with the truth as Plato’s Socrates envisaged, because it lacks the active—that is, the interpersonal—ingredient that turns, as if by alchemy, raw information into personal insight.

Of course, the internet itself does not necessarily entail the end of wisdom. It remains possible, after all, to use its resources whilst still practising an embodied form of learning, just as it is possible to bring a book to life by intentionally extending quasi-personhood to it—a power unique to persons. Indeed, it is often said that reaching the end of a good book is like saying goodbye to a close friend. And books can be great aids to communal learning and ongoing discussion; the dead letter can be resuscitated by the living teacher. Whilst understanding information should not be mistaken for wisdom itself, the “reminder” has its place, and can still engage the rational soul and lead it to truth if properly mediated. This, in fact, is how a great teacher uses books with his or her students.

In the case of AI, however, we are dealing with something far more dangerous—for AI effectively removes the need for human mediation altogether. Up until very recently, the process of converting raw information into, for example, an essay, would have still required us to engage our faculties of memory, inference, and reason, and generally would have benefited from discussion with a teacher or peer. These processes in turn deepen the mind’s aptitude for wisdom. But with the rise of technologies like ChatGPT, it is now possible to outsource cognition to an inhuman entity. This brings us perhaps to the most serious threat posed by AI, at least from a Platonic perspective: the subcontracting of our mental capacities, and in particular memory—an essential faculty for wisdom.

Recall how, in the Phaedrus, the Egyptian king noted that the problem with writing is that it removes the need for human beings to remember things from within themselves. This is because writing acts as an artificial prosthesis that retains information on our behalf. As discussed, the internet exacerbates this; yet, even with the internet, we still must internalise information for ourselves when going about any scholarly or creative endeavour. AI, however, makes it possible to circumvent those processes altogether. We can get bots to take lecture notes for us without us even having to register what is being taught; and we can conjure up a comprehensive essay about a subject without having absorbed the relevant information ourselves or having had any personal insight. In other words, we can go about the pursuit of knowledge without internalising anything.

In another of his dialogues, the Theaetetus, Plato reveals why the internalisation of knowledge is so essential to wisdom. He compares the faculty of memory to a “block of wax,” onto which our sense impressions are stamped, and posits a positive correlation between the strength of our wax and our overall intelligence. This is because, the stronger our wax, the more distinct the imprints of things in the external world, which can subsequently be “quickly assigned to their several stamps—the ‘real things’ as they are called.” In other words, those with strong mental faculties are able to arrive at the knowledge of true forms; and so, he tells us, “are said to be clever.” Those whose wax is soft, on the other hand, are forgetful; not only do they struggle to retain impressions, but they also, subsequently, struggle to realise the true nature of things.

Thus, memory—in allowing us to internalise sense impressions and form abstract associations between them—is what first enables our ascent towards truth. This Platonic account profoundly influenced the educationalists of the Middle Ages, who interpreted it to mean that we should actively work to strengthen our memory as if it were a muscle. This was the purpose of the medieval memory arts, like those devised by Ramon Llull and Giulio Camillo, described by Frances Yates in her famous book The Art of Memory. Therein, Yates surveyed how the improvement of memory was considered an essential part of pedagogy—and how anything that weakened it was strongly condemned. It was condemned because anything which inhibits memory, including that which remembers for us, softens our wax: it renders us less capable of relating the objects of our senses to the true nature of things and so, ultimately, inhibits wisdom.

We can now understand why, from a Platonic perspective, the exteriorisation of knowledge to inhuman prostheses is so inimical to the pursuit of wisdom: not only does it remove the interpersonal element which is needed to guide each individual soul to truth, but it also weakens the very faculty that allows us to internalise information and transform it into knowledge and insight in the first place.

Especially when it comes to education, then, we should be worried about AI. Not only does it encourage intellectual laziness and plagiarism, but it fundamentally undermines those aspects of our very human nature on which the acquisition of real knowledge depends. With AI, we have not only an exaggeration of the Wellsian World Brain—a disembodied mass of information capable only of reconfiguring dead letters, lacking the human mediation needed to bring them to life—but a new prosthesis that exteriorises memory and so diminishes our ability to arrive at truth ourselves. In essence, AI has the potential to take the course of de-personalising our existence to its final extrapolation.

As is often the case in the modern world, many will claim that this radical cutting of corners is a mark of progress; a means for us to get on with more worthwhile (by which they usually mean more profitable) pursuits. But what they do not realise is that in undermining our very human nature, and neglecting our unique ability to seek and come into union with truth, AI will leave us not only uninspired but undignified. To echo King Thamus’ prophetic remarks once again, it will implant forgetfulness in our souls.