AI and the Cult of Intelligence

A narrow notion of intelligence fuels both worship and fear of AI

By Paul O’Connor

Artificial intelligence has seized the collective imagination, as an object of both utopian dreams and apocalyptic nightmares. Much of the reason lies in the name. If ChatGPT and its equivalents were described as ‘automated word predictors’ or ‘text completion tools’, which is arguably what they are, they would not exert such a powerful hold on the cultural imaginary.

However, ‘intelligence’ – a very limited notion of intelligence, at that – has a central place in the self-definition of modern humans. It is because we idolise a particular kind of (disembodied, abstract, ratiocinative) intelligence that advocates of AI have persuaded themselves that the technology can solve all the world’s problems, producing revolutionary breakthroughs in medicine, halting climate change, and ushering in a post-scarcity society. At the same time, it is because we moderns have come to identify so strongly with the same faculty that we fear ‘superintelligent’ AI could potentially replace us.

Both hopes and fears are probably exaggerated, in part because ratiocinative intelligence – artificial or otherwise – is not all it’s cracked up to be. We humans are not defective, unevolved computers but embodied beings who live on this earth and interact with our surroundings sensuously and emotionally as well as cognitively, as thinkers from Gregory Bateson to Martha Nussbaum and Tim Ingold have argued in depth. Those cognitive processes validated as intelligence are only a tiny aspect of the full spectrum of our ‘being in the world’, and furthermore are inextricably intertwined with and powered by emotion and sensation.

Could a disembodied data processor – no matter how good at crunching numbers or predicting text – ever be truly intelligent? More fundamentally, what is the consequence of separating ratiocination from embodied human experience, emotion and empathy? If such an ‘intelligence’ were to develop autonomously, and evolve some form of consciousness and will, the likely result would be something warped and insane - not the god in the machine, but a demon whispering in the crawlspaces of the web.

Is AI Intelligent?

As is well known, the current crop of AIs like ChatGPT and Deepseek operate on the basis of what are termed ‘Large Language Models’ (LLM) of machine learning. Basically, vast quantities of information harvested online are fed into the algorithm, which learns to recognise repeated language patterns and word associations, and replicates these in response to the questions put to it. In addition, the AI is trained to respond to instructions, and receives what is known as ‘reinforcement learning’ through human feedback. An excellent explanation of how these AIs work is available here: https://medium.com/data-science-at-microsoft/how-large-language-models-work-91c362f5b78f. The effectiveness of the model is basically a function of the amount of data available to it and the computing power at its disposal. Future advances in AI are expected to come about through increasing both these factors.

But does this kind of automated pattern recognition really deserve the name ‘intelligence’? Okay, so it can generate a 1500-word essay on virtually any topic you choose within a couple of minutes, but it is summarising and mimicking text already available online. Is generative AI not parasitic on the products of genuine human intelligence? This is certainly the position of the many authors, artists, musicians and other creators who are protesting the unpaid appropriation and exploitation of their work by AI tools.

AI models based on LLMs can summarise and rearrange the material they assess with incredible rapidity and growing accuracy to answer questions and respond to prompts. But unlike a human, they cannot experience, reflect on that experience, and so create something genuinely new.

If generative AI is parasitic on the products of human culture, ought we really be celebrating it as some kind of great advance? As people become dependent on AI for producing everything from college essays to newspaper articles, film scripts to laws (the UAE is now using AI to write legislation), don’t we risk entrapment in a closed loop of endlessly rearranged text, producing a sophistic culture ever-more detached from concrete experience or genuine self-expression?

Intelligence and Experience

We tend to associate intelligence with ratiocination or problem-solving, and define this in narrowly mechanistic terms: if sufficient ‘high-quality’ data is fed in, and enough computing power applied, the assumption is that the correct solution will automatically be spat out. This ignores the fact that the definition of something as a ‘problem’, and the selection among possible ‘solutions’ are matters of judgement inseparable from experience.

When I use the term ‘experience’ it is in the sense of the German philosopher Wilhelm Dilthey, whose concept of ‘lived experience’ (Erlebnis) sought to take account of ‘the whole of human nature in which willing, feeling and thinking are only different aspects of the real process of life’ (Dilthey, 1976). What Dilthey argued for was a kind of empiricism, but one very different from that proffered by either the British empiricists or Immanuel Kant. As he writes, ‘epistemology has, up till now, explained experience and cognition merely from the facts of apprehension. No real blood flows in the veins of the knowing subject constructed by Locke, Hume, and Kant; it is merely the diluted juice of reason, a mere process of thought’ (Dilthey, 1976: 162). By contrast, Dilthey outlines an inclusive empiricism which seeks to capture the lived reality of experience as it presents itself to our consciousness. We know the world through our feelings and strivings as well as through our sense impressions and thinking, and we know it from within a particular individual, social and historical context. Knowledge of the world develops through our everyday life and our involvement with our environment, and is inseparable from this.

Conceptual thought captures just a tiny fraction of the bandwidth of human experience. It is a useful tool for reflecting upon and sharing this experience, but to mistake it for reality is like arguing that a map is more real than the territory it represents. The reification of concepts – treating them like things, constructing a hallucinatory world of abstractions piled upon abstractions – is a characteristic disease of contemporary thought. AI will only exacerbate this tendency. The only way to truly get to know a place is by walking around it, but for AI nothing exists beyond the map. It is the ultimate idealist, for it will never kick a stone. In fact for AI a stone can only ever be a string of words – ‘hard solid non-metallic mineral matter’.

No doubt AI will continue to evolve, in speed, scope, and the ability to ‘autonomously’ learn and solve problems. But it is hard to imagine an algorithm having an ‘experience’. And if AI cannot experience, can it ever be truly intelligent or creative?

Dickens and the Reasons of the Heart

Not only can a machine not have an experience, neither can it possess empathy – and arguably empathy is a crucial aspect of genuinely human intelligence.

One way to illustrate this point is with reference to the selections from Hard Times published in this Substack a few weeks ago. Thomas Gradgrind is undoubtedly what Donald Trump would call ‘a high IQ individual’ (see https://nevadacurrent.com/2025/02/16/elon-musk-brought-his-4-year-old-to-the-oval-office-it-wasnt-just-a-photo-op/). Sissy Jupe on the other hand displays a distressing inability to get her head around basic ‘facts’. Yet Gradgrind’s blindness and – one can only call it this – stupidity destroys his family and ruins the life of his favourite child, while Sissy Jupe displays not only more true understanding, but the moral authority and force of character to overawe the urbane and wealthy James Harthouse and persuade him to give up his pursuit of Louisa.

Dickens demonstrates that there are reasons of the heart as well as of the mind, and the former are indeed more fundamental. Note that Harthouse terms Gradgrind, contemptuously, ‘a machine’. The utilitarian thinkers of the mid-nineteenth century, like their intellectual descendants in Silicon Valley today, believed ‘progress’ and social improvement would invariably follow from the operation of a reductive, mechanical style of thought, for which a horse is a graminivorous quadruped with forty teeth, and that’s all there is to it. Hard Times shows the deathly effect of such ‘intelligence’ as only a truly great novel can, and its lesson is more relevant than ever today.

One cannot divorce intelligence from embodiment and imagination and emotion. Ratiocination – a favoured word of both Kant and Habermas – separated from any of these things is not intelligence: it is at best defective and can easily become monstrous. Ratiocination separated from empathy is particularly dangerous, because it is blind to human motivations and treats people as objects rather than autonomous and feeling beings.

Here again we find the essential connection between human intelligence and embodiment. We feel empathy because, having had experience of physical or emotional pain ourselves, we are able to imagine the pain of others. A reasoning machine is incapable of empathy. Insofar as we humans outsource our thinking to machines, or attempt to model our thinking on machines, we too lose the capacity for empathy and become less, rather than more, intelligent.

An Impoverished Notion of Intelligence

Behind all this lies the fact that while modernity worships intelligence, it has a strangely impoverished notion of what it is. This feeds directly into both contemporary celebrations of, and fears about, AI. The dominant conception of intelligence views it as disembodied, ratiocinative, quantifiable, instrumental and separable from context, morality, or emotion. It is exemplified in the concept of ‘IQ’ – the belief that intelligence is a single variable and can be relegated to a specific number derived from a standardised test. This is not only reductive, but culturally contingent to modern, westernised societies.

By way of contrast, it is interesting to look at some of the terms used for those with great intellectual abilities in the past, and the type of connotations these had.

Take the Romantic conception of ‘genius’. Today the word genius calls to mind a scientist like Einstein or a kid scribbling equations on a blackboard (do a Google image search and see), but in the nineteenth century the archetypal genius was a great artist like Leonardo or Rembrandt, or a poet like Shakespeare or Goethe. Genius connoted not only (or even primarily) ratiocinative ability but creativity, force of character, and a kind of greatness of spirit that could magnetise the wills of others and shape the history of nations or humanity as a whole. Echoing its original meaning in the ancient world, it evoked the idea of some force beyond purely rational, human explanation.

Or take the term ‘savant’, which came into broad use in French and English in the early modern period, and remained common through the nineteenth and early twentieth centuries. Despite its association with the Enlightenment, the connotations here are not of purely abstract intelligence, but possession of a high degree of learning and skill, especially in some specific field of knowledge or art.

The most fundamental notion, and one with cognates in every human culture, is wisdom. Wisdom, as opposed to mere cleverness or technical know-how, involves knowing how to live well, act rightly, and maintain social and cosmic harmony. Moreover, wisdom is inseparable from experience. The archetype here is the ‘wise old man’ or ‘wise old woman’ – somebody who has lived a long time, experienced much, learned from it, and so can give guidance and advice to others. Such wisdom is literally embodied, weighing on their shoulders or engraved in the furrows of their brow.

It is noteworthy that practically every culture distinguished wisdom from mere cleverness, cunning, or technical knowledge. The latter was viewed as inferior, even dangerous, being associated with trickster figures such as Anansi, Coyote, Loki or Prometheus. Modern culture is unusual in uncritically celebrating, even deifying, abstract ratiocinative intelligence while ignoring wisdom – or even outright mocking and subverting it.

The Descent of Reason

There is no space here to construct a detailed genealogy of our contemporary notion of intelligence, but it is possible to identify and peel back a few layers.

One is the invention and dissemination of standardised IQ testing. The first IQ test was developed in 1905, by French psychologist Alfred Binet, specifically to identify ‘mental retardation’ in school children, while the practice of IQ testing was popularised by its use in U.S. military selection during WWI – hence its origins and spread are tightly wedded to practices of governmentality. It is to the popularisation IQ testing we owe the idea that intelligence can be reduced to a number, while it also reinforced the association of intelligence with a rather narrow set of cognitive abilities (mainly verbal and mathematical), focusing not on the memory of lived experience, the basis of wisdom, but on the regurgitation of a series of numbers or words, excluding practical, social or emotional intelligence.

Reinforcing these associations with governmentality, it is significant that ‘intelligence’ is the term used for state agencies such as the CIA or MI6 dedicated to gathering and analysing information, especially secret information, about other countries, and even their own citizens. The detective, the spy, and the secret policeman are all avatars of modern intelligence; but so too is their schismogenic double, the revolutionary intellectual. ‘Intelligentsia’, originally a Russian word, was the self-descriptor of the critical intellectuals, ‘woke’ to every contemporary injustice, who were analysed by Dostoevsky in Demons and whose descendants provided the leadership of Lenin’s Bolsheviks.

Another layer in any genealogy of ‘intelligence’ is the development of the modern research university in nineteenth century Germany. This is important because the German universities, which became the model for American research universities and so indirectly for universities throughout the world, were deeply permeated by NeoKantianism. Reviving and transforming Immanuel Kant’s critical philosophy, NeoKantianism was characterised by its conviction that knowledge is actively constructed through conceptual frameworks, and science advances through the refinement of concepts. This attitude is partly responsible for the modern tendency towards reification, or the treatment of concepts as if they were real things, which in turn leads to a simplistic identification of ‘intelligence’ with a facility for purely performing narrowly conceptual operations.

Yet another layer in the evolution of the modern concept of intelligence was the Cartesian separation of mind from body, and therefore of intellectual from sensuous experience. Without this separation, it would be impossible to even conceive of ‘artificial intelligence’ as a possibility: the phrase would be a self-evident oxymoron.

Finally, the Baconian reduction of knowledge to operation, and dismissal of the quest for meaning in favour of the accumulation of human power over nature – a trickster revolution which enacted a truly catastrophic ‘revaluation of values’ – was crucial to the contemporary confusion of intelligence with instrumental or substitutive rationality.

What we see in each case is a narrowing and reduction of how our intellectual faculties are understood. Perhaps most fundamentally, they are no longer viewed as faculties whose most important purpose is the discernment of a meaningful order – whether natural and cosmic, or human and social – which is already given to us, but as tools to be deployed in the construction of artificial structures of knowledge and social organisation which are instrumental and contingent. We have ‘progressed’ from contemplation of Ἁγία Σοφία (‘Holy Wisdom’, from whence derived the name of the cathedral church of Constantinople, now a mosque) to the mindless idolatry of smart objects.

This is the context within which the advent of AI exerts such a magnetic power upon the contemporary imagination. If intelligence is mere information-processing power or logical reasoning, machines can appear either awe-inspiring or threatening precisely because they can outperform us at these tasks. If we broaden our idea of intelligence, we will be better able to better judge the true potentials, and dangers, of AI for both society and our own human nature.

Reference

Dilthey, Wilhelm (1976) Selected Writings (transl. H. P. Rickman), Cambridge: Cambridge University Press.