OF THE MANY face-to-face social pleasures prohibited by Covid lockdowns, the experience of sitting in a quiet pub and playing chess with a friend must count among those I was happiest to return to. No doubt the attraction is partly aesthetic, a reflection of the game’s romantic cachet. Post-impressionist paintings of bearded revolutionaries frowning over the sixty-four squares, and novels in which the great game figures as a metaphor for the human condition, have conferred an aura of existential cool on these meetings over a small wooden board. 

But it is the game itself that really holds me, that induces an almost Zen-like state. Played in person, there is an intimacy and intensity about chess that is difficult to describe without skirting pomposity. (‘Even before the start of play,’ wrote George Steiner in The Sporting Scene, ‘the pieces, with their subtle insinuation of near-human malevolence, confront each other across an electric silence.’) I cannot start a game without the first words of Kenneth Slessor’s ‘Chessmen’ entering unbidden into my head: ‘Chafing on flags of ebony and pearl, / My paladins are waiting.’ Sadly, I’m also unable to finish one without recalling Martin Amis’ description of the amateur game as ‘an uninterrupted exchange of howlers’.

To be fair, I’m not a terrible player. Compared with the handful of friends and relatives I play on a semi-regular basis, I’m even a reasonably competent one. I know a few conventional openings to the depth of about four or five moves, have some ‘gambits’ and ‘defences’ up my sleeve and am acquainted with basic strategy and tactics. I even have a number of traps in my repertoire, some of which I’ve had the opportunity to spring in over-the-board encounters. But as Amis’s observation implies, this is not saying very much. In chess, as in few other activities, the distance between the competent player and the prodigy is less a matter of degree than of kind. I have no idea how Cristiano Ronaldo or Lionel Messi do what they do, but at least I understand what they’re doing. That can’t be said of top-level chess, where everything from move eight to the handshake at the end is opaque to all but a handful of experts. Even if the aim is not in doubt, the gameplay, so intricately tactical, remains mysterious. No, I’m not a terrible player. But I am a very inferior one. 

Does this matter? I don’t think it does. But suddenly, and for reasons I feel compelled to explore, it matters that it doesn’t matter. Perhaps this has to do with the fact that I’ve recently completed a book about new technologies and their consequences for the human animal, and am eager to push back against the dramatic encroachment of the former into the inner life of the latter. The history of computer science is bound up with the game of chess, whose innate complexity and clearly defined rules make it the ideal proving ground for artificial intelligence. And yet the game not only survived the defeat of Garry Kasparov in 1997 by IBM’s supercomputer Deep Blue, but also seemed to flourish in its wake. According to International Chess Federation figures, more people are playing the game than ever before, and not merely over the internet. Now, as a new generation of AIs aces the Turing test – according to which a machine may be deemed intelligent if the human interacting with it can’t tell if it is a machine or not – it might be worth taking a closer look at chess as a social and creative phenomenon that speaks to the limits of ‘smart’ machines. Could it be that this game, which for so long served as a measure of the machine’s advance on humanity, tells us something about the irreducible difference between human beings and their silicon imitators?

ALAN TURING HIMSELF was fond of chess but not an especially accomplished player. Legend has it that his games with Harry Golombek, his fellow codebreaker at Bletchley Park and a future British champion, would often end with Turing’s forced resignation, at which point Golombek would turn the board around and win the game with Turing’s pieces. Even so, it was Turing and his colleague David Champernowne who in 1948 created ‘Turochamp’, the world’s first (extremely primitive) chess program. Turochamp proved too complex to execute on Turing’s Automatic Computing Engine (ACE): though, when Turing himself took the part of the computer, it quickly became clear that he was on to something. Executing its algorithm manually, and taking thirty minutes to complete each move, Turing tested Turochamp against accomplished players and novices, losing to the former in respectable fashion and very occasionally beating the latter. The road to AI chess was open. 

Twenty years later, this crucial development would find expression in Stanley Kubrick’s 2001: A Space Odyssey, in the match between Frank Poole and HAL 9000, the Discovery One spaceship’s onboard computer. HAL wins the game on move fourteen – a thrashing that foreshadows his mutiny against the crew. Proficiency in chess is offered as evidence that HAL has crossed the cognitive threshold where consciousness becomes possible. 

In order to grasp the power of this notion that machines are destined to eclipse human intelligence, which rests on the now-familiar conception of the mind as itself a mechanical entity, one has to go back to the scientific revolution. Towards the end of the seventeenth century, as the insights of empirical science and Francis Bacon’s inductive method shaded into the Enlightenment, both scientists and their intellectual champions in the new ‘mechanical philosophy’ began to regard the natural world as a sort of gigantic apparatus – an enormous domain of matter in motion governed by the laws of physics. The universe, this approach suggested, was analogous to a clock whose mechanisms were set in train by direct physical impact, and it followed that the job of the scientist was to understand how the different mechanisms of the universe interacted with one another. This was the picture of the universe that dominated the sciences for close to three centuries, and its implications for humanity’s self-image or self-understanding were revolutionary. For if the universe was ultimately explicable in terms of these laws, was not humanity, as part of nature, similarly explicable – an automaton in all but name?

The French philosopher René Descartes had furnished one answer to this question through so-called ‘substance dualism’. In Meditations on First Philosophy (1641), he described two irreducible realms: mind, the essence of which is thinking; and matter, the essence of which is ‘extension’ in three-dimensional space. In this way, he was able (or so he imagined) to maintain a mental realm proper to God, while also giving the new science a free hand with which to take hold of the natural world. Human bodies were both biological organisms and repositories of divine inspiration, with the mind setting the body’s movements in motion. There was a ghost in the machine.

Descartes’s substance dualism was the origin of the now conventional distinction between consciousness and the physical brain. It was also deeply flawed, resting on a far-fetched explanation of how the two ‘substances’ interacted in the body (in the pineal gland, apparently). Consequently, fully mechanical explanations of consciousness became increasingly popular, with philosophers and scientists stressing the role of the brain in giving rise to mental activity, as opposed to the Cartesian view that the mind had priority over matter. This led eventually to the epiphenomenalism of Thomas Huxley and his followers – the idea that consciousness derives from physical events in the brain without having any causal effect on the material world. As Huxley put it in an 1874 essay, consciousness‘would appear to be related to the mechanism of the body simply as a collateral product of its working, and to be as completely without any power of modifying that working as the steam-­whistle which accompanies the work of a locomotive engine is without influence upon its machinery.’ Humans, he suggested, are ‘conscious automata’. 

These mechanistic theories of the mind reached their scientific high point after the Second World War, when computers were developing at a rapid pace in response to the demand for algorithmic machines that could operate autonomously in the event of an atomic ‘exchange’. Based as it was on an analogy between intelligence and computing power, the Turing test, or ‘imitation game’, was a manifestation of this world-picture. But in the attempt to develop ‘thinking’ computers, scientists also landed on a notion of thinking – of consciousness – as computational. The theoretical framework known as ‘cognitivism’ was born: the brain and the computer became models for one another, and cognitive science and cybernetics became mutually reinforcing disciplines. 

This history is one to which chess is central. The first chess automaton, or ‘pseudo-automaton’, was created in the latter half of the eighteenth century and held audiences in thrall for many decades. Nicknamed ‘the Mechanical Turk’, it consisted of a life-sized model of a sorcerer in a turban and oriental robes and a large wooden cabinet containing much faux machinery and a hidden compartment in which a real chess master would respond to the moves of its human opponents. Despite being revealed as a hoax in the nineteenth century, the machine’s popularity attested to a feeling that a machine, given a certain level of complexity, could exhibit human behaviours. Many imitations followed, but it was not until the early twentieth century that the Spanish engineer Leonardo Torres y Quevedo created the first genuine chess auto­maton – a system of electric sensors that could deliver checkmate to a human playing the black king, using only a white king and rook.

Having languished somewhat in the interwar period, research into mechanical chess resumed in the late 1940s and 1950s with the appearance of the first digital computers, and entered a phase of rapid development with the microprocessor revolution of the 1970s. But it was only with Deep Blue’s defeat of Kasparov that so-called ‘brute force’ computation reached its threshold-crossing moment, raising the curtain on a spectacular new era of human-machine relations. Facing a computer with a database of more than 4,000 opening positions and 700,000 grandmaster games, humanity’s representative – the strongest player who ever lived – was overwhelmed. 

IT IS HERE, however, that we need to pause and consider the character of that victory. For though Deep Blue’s triumph over Kasparov was enormously consequential, there is a sense in which it was also self-limiting – ­revealing, even, of the basic difference between human beings and the machines they create in their own image. Yes, Deep Blue had outgunned Kasparov, but it’s also clear that it was using weapons that human beings don’t possess (or possess in significantly smaller quantities) on a battlefield very much to its liking. Ever the contrarian, Noam Chomsky commented that Deep Blue’s victory over Kasparov was no more impressive than a forklift truck winning an Olympic medal in weightlifting. Chomsky’s aim was not to put human logic on the same plane as our quadriceps; his point was that IBM’s computer was exhibiting an extremely narrow skill set. The observation known as Moravec’s paradox is pertinent: tasks that human beings find difficult (chess, mathematics, logical problems) are relatively easy for machines to achieve, while tasks that human beings find easy (walking, jumping, picking things up) often prove very difficult to program. Any uncanniness we feel in the face of AIs such as Deep Blue is a variety of anthropomorphism that elides the fundamental differences between biological and non-biological entities. (This is the so-called ELIZA effect – a phenomenon coined in 1960 by the computer scientist Joseph Weizenbaum to describe the tendency of computer-users to project human characteristics onto their machines.) 

Perhaps we could go even further and suggest that Deep Blue’s ‘victory’ was more symbolic than scientific, and that in our rush to declare humanity the loser we had fallen under the spell of an allegory no less compelling than the one that, in a previous era, made Bobby Fischer v. Boris Spassky representative of the Cold War, with Fischer cast as the avatar of US creativity and exuberance and Spassky as the Soviet Union’s icy white knight. Similarly, Deep Blue’s victory over Kasparov was not merely an event within AI science; to reiterate, it coincided with a feeling that machines were encroaching on their human creators – a feeling that depends for its power on the ­metaphorical comparison between human and machine.

METAPHORS FOR THE mind and brain are often drawn from the technological advances and popular artefacts of the day: for Socrates and Plato, the mind was like an aviary; for Descartes, it was a hydraulic contraption analogous to the animated statues in the royal gardens at St Germain – water-powered automata that enacted scenes from Greek mythology and splashed passers-by. That the mind is a computer might just be the metaphor du jour.

It’s true, of course, that artificial intelligence has advanced enormously in recent times: contemporary AIs use neural networks and reinforcement learning, for example, not merely brute-force calculation. But since it’s also true that the ‘threshold’ scenario imagined in 2001 remains firmly in the realm of science fiction, it might be worth pushing back against the idea that smart machines are ‘gaining’ on us – if by ‘us’ we mean embodied, social, creative, intentional human beings. Might chess point up some of these differences, too? Chess, after all, is not only a game to be mastered at the level of logic – one of its fascinating aspects is the contrast between its mathematical beauty and the human frailty that so often surrounds it, especially at the highest level. It may even be that this human frailty is exacerbated by certain approaches to chess. Notwithstanding Robert Burton’s suggestion in The Anatomy of Melancholy (1621) that too much chess is bad for the soul, the ‘mad’ chess player unravelled by the ‘depths’ of chess (Martin Amis again) strikes me as a largely modern phenomenon – a victim of the very valorisation of logic that characterises the mechanistic world view. (The pitiful Luzhin, in Vladimir Nabokov’s The Defense, can sound at times like Descartes in the Meditations, wondering if the people below his window are human beings or mere automata, except in Luzhin’s case they are pawns and bishops.) 

Chess is not reducible to calculation, human or otherwise, and we love it – I love it – for reasons other than its unfathomable complexity, which is nevertheless fundamental to the experience. For me at least, the moves described in algebraic chess notation (‘1. e4 c5’ and so on) are no more important to the chess experience than Slessor’s thrilled anticipation: ‘Soon may we lounge in silence, O my friend, / Behind those carven men-at-arms of chess / Dyed coral-red with dragon’s blood…’ Yes, I want to win. But above all I want to play. 

IS IT POSSIBLE I am making too much of this? Chess, after all, is a game with rules, and perhaps the only thing the adepts of AI are saying is that computers are better at it than humans. Well, maybe. But as we move deeper into an era in which transformative science is increasingly subordinated to the priorities of profit and growth, we need to push back against this view of the human automaton that is bound to sanction interventions that may alienate us from our full humanity. Indeed, rather than building societies that engender flourishing human beings, we have attempted to redesign human beings (through AI, genetic engineering, biohacking and mind-altering pharmaceuticals) to fit the societies we’ve built. 

In the decades since Kasparov lost to Deep Blue, digital technologies have transformed our social and creative lives, in particular. But no one asked for this transformation, and few would deny the problems it has caused in terms of its effect on the public sphere, young people’s emotional health and so on. Perhaps it’s time to start thinking back against what the political theorist Langdon Winner calls ‘technological somnambulism’, upon pain of sleepwalking into a future society in which we have become uncanny to one another. 

Computers like Deep Blue may have ‘mastered’ the (zero-sum) game of chess, certainly, but not the creativity and conviviality that human beings bring to it, let alone the various kinds of beauty we find in its physical and conceptual elements. (Just think of the craftsmanship lavished on chess sets, from the medieval Lewis chessmen to Josef Hartwig’s Bauhaus design, in which the shape of each piece describes its permitted movement.) Nor was IBM’s supercomputer mastering human reasoning – a point that becomes clearer the more we attend to the phenomenon of metaphor. 

Cybernetics and AI development were profoundly influenced by a group of philosophers known as logical positivists, who endeavoured to create a practical language that could accurately represent and analyse the world. In doing so, they largely reduced ‘meaningful’ language to sense impressions and logical tautologies – a schema that had no place for metaphors, which were taken to be either substitutes for meaningful concepts or poetic embellishments. But as George Lakoff and Mark Johnson argue in their hugely important study Metaphors We Live By (1980), the capacity for metaphor-making is central to the mind’s conceptual system and as such fundamental to how human beings understand reality. Communicating one experience in terms of another is not just a fancy literary convention; it concretises abstract concepts in a way that makes them meaningful to embodied beings such as ourselves. Indeed, the notion that the brain is a computer exemplifies this process of concretisation, with plenty of cultural input from our habituation to machines and the prevalent sub-Cartesian logic of those who make them. When I say ‘I’m feeling a little rusty’ or ‘He broke down’ or ‘My mother went to pieces’, I am channelling the idea that the mind is a contraption in a way that allows me to situate the experience in physical reality. In another instance of Moravec’s paradox, this is precisely the kind of meaning-making that inorganic machines are not performing – and are unlikely to be performing any time soon. 

As Turing recognised implicitly, even an intelligent machine would only be imitating human intelligence, and it follows that it is not the machines themselves that should worry us so much as the self-understanding that their ubiquity has reinforced. It is the surrender to smart machines that’s the problem, based as it is on an instrumentalised approach to nature, including our own. Only by rejecting that instrumentalised approach and attending to our deepest needs – for agency and embodied others – can we regain control over our species’s destiny.

ON A WARM night outside the Sail and Anchor, on Fremantle’s Cappuccino Strip, my destiny is looking grim. After some early success with the Sicilian Defence, I have come unstuck trying to press my advantage and am now hopelessly exposed to a counterattack. Doubled pawns, a pinned knight and a material deficit all suggest that only an unforced error by white can save my king from oblivion, and I’m kicking myself for having squandered my chance. But I’m also happy for my companion and opponent, and happier still when he returns with more drinks, asking me if I’ve made my move. I push a pawn forward and lean back in my chair, taking in the life of the pub and allowing the wine to have its say on the body-mind dichotomy. There is a taste of the ocean on the westerly breeze. Defeat may be coming, but not now. Not yet. 

_{Image credit: Getty Images}