It has long been the prevailing wisdom that the humanities are in crisis; that in a world in which rational process and technocratic prowess are at a premium, anyone with any sense should be studying a Stem subject, not history or literature. It’s a fashion that has seized the glibbest of minds: those who once told us that globalisation was an unstoppable force now assure us that the algorithm is all and that AI is set fair to transform the world.
I’m not so sure. “Cui bono?” was the question Thomas Hobbes urged us to ask the metaphysicians—and surely the fad for so-called artificial general intelligence is the latest metaphysical irruption. Current crazes are feeding the wealth of those who promote them without concern for the apocalypse they claim to foresee or for the crash that unbridled investor enthusiasm so often brings in its wake. If the AI bros really think they are creating an existential threat, why are they so gaudily gung-ho for annihilation? The share price, silly. Invocations of doom are, with exquisite perversity, just another ingredient in the recipe for hysteria or what Charles Mackay, back in the 1840s, called “Popular Delusions and the Madness of Crowds”.
Just over 10 years ago, I was one of the visiting Humanitas professors at Oxford, funded by the visionary publisher George Weidenfeld—an early scheme to push back against the perceived marginalisation of the humanities. In my inaugural lecture, I used economist John Kay’s subversive notion of obliquity to try and explain why the humanities are not only one of the ends of any civilised life—the unexamined life, as Plato’s Socrates had it, is not worth living—but also a means of harnessing the unexpected to the apparently philistine juggernaut of progress. Kay was inspired to write a book about obliquity by a remark of Sir James Black, the Nobel prize winning pharmacologist, to the effect that “you are often most successful in achieving something when you are trying to do something else”.
The humanities, being human, can point one in unexpected directions. They evade the algorithm. It’s why they should be at the centre of any university worth the name.
The dangers of modern life and its social order were identified more than 100 years ago by the German sociological pioneer Max Weber. What he called the disenchantment of the world goes hand in hand with the tightening grip of an iron cage of rationality (Talcott Parsons’s compelling translation of Weber’s original stahlhartes Gehäuse or “steel-hard casing”). The burgeoning vogue for Rory Sutherland, the self-styled Wikiman, writing in the Spectator and popping up all over the internet, stems from a shared sense that something is wrong with this reign of overweening reason. Sutherland’s point is that adherence to process, the fulfilment of targets and the pursuit of the ultimate rational goal, efficiency—these all end with us pursuing formal rather than substantive reason. The kingdom of the accountants and the HR department squeezes out innovation and imagination.
In the early 1990s, before I became a singer, I worked for one of those little independent TV companies that sprang up to service the needs of Channel 4 and the new commissioning regime at the BBC. I was mostly working on business and politics programmes—an interview series with the fabulously eccentric Mary Goldring, ex-Economist, presenter of Analysis on Radio 4 and one-time scourge of supersonic air travel—but we also made things for a science strand, Equinox.
The large language models we are all so excited by are not a royal road to anything one might call ‘consciousness’ or even ‘intelligence’
My contribution was to make a film of Roger Penrose’s 1989 book, The Emperor’s New Mind: Concerning Computers, Minds and the Laws of Physics. It left me with a deep suspicion of futurists such as Ray Kurzweil, an early proponent of uploading minds onto computers, unhealthily excited by the possibility of acquiring a new shiny metallic body. The latest prophecies of an imminent “singularity” on the back of the impending arrival of AGI are surely desperately premature, if not utterly misguided. This isn’t the place for philosophical amateur hour, but the large language models we are all so excited by are not a royal road to anything one might call “consciousness” or even “intelligence”. They are very sophisticated imitation machines and indecently benefit from the stubborn human propensity to anthropomorphise. Remember, way back when, those apparently effective psychotherapy computer programs with their mechanistic patter: “How do you feel?”, “I know what you mean”, “Tell me more…”. The famous Turing test for machine consciousness—can a computer trick you into believing that a conversation with it is real?—sets an exceptionally low bar, and the sexed-up LLM version isn’t, in principle, much different.
Penrose’s more subtle, if highly disputed, point was that human thinking, and in particular mathematical insight, is non-algorithmic; an argument based on Kurt Gödel’s famous incompleteness theorem (in any consistent formal mathematical system, there are true statements that cannot be proven within the system itself). Which is to say, not everything can be reduced to code.
In an age in which our humanity is under attack from so many sides—the inexorable march of bureaucratic rationalism allied to the efficiency-driven march to AI—the humanities are more important than ever. Because they remind us what it is to be human.
