Back in 2016, a month before the EU referendum, I went along to the Future of Humanity Institute in Oxford to interview its director, the Swedish-born philosopher Nick Bostrom, who had just written a book called Superintelligence. The book outlined the existential risk to democracy and humanity implied by advances in machine learning. Bostrom’s institute, which sought to weigh the apocalyptic potential of various humanity-threatening forces, had just been given a £1m grant by Elon Musk.
If I think about that encounter, I remember three things. The first was that when I arrived a bed was being delivered to the institute, cementing the belief that anxiety about impending catastrophe was, these days, a 24/7 kind of occupation. The second was that the germophobic Bostrom was the first interviewee I’d met who insisted on fist bumps rather than handshakes (the shape of things to come). And the third was his insistence that in important ways, artificial intelligence posed a more imminent threat to the survival of our species than, say, climate crisis or pandemics or nuclear war.
At the time, that idea seemed to me inflected with too many outlandish tropes from science fiction. In the seven febrile years since, less so. Bostrom’s book presented several scenarios in which our fate may be sealed by the machines we create. One projection involved an AI system building covert “nanofactories producing target-seeking mosquito-like robots [which] might then burgeon forth simultaneously from every square metre of the globe” in order to destroy us. Another, somewhat more credible vision saw a superintelligence “hijacking political processes, subtly manipulating financial markets, biasing information flows” to bring about first our superfluity, then our extinction.
David Runciman’s far more sober book is in part an analysis of the first phase of that latter proposition. The Cambridge politics professor and, until it ended last year, the ever-erudite host of the terrific Talking Politics podcast is not given to apocalyptic prophecy. The closest he came to it was a series of ironic exclamation marks that revved up the chapter headings in his 2018 book How Democracy Ends. One of those was devoted to “Technological Takeover!”, which, with parliament then in Brexit meltdown and Trump in full spate, examined the ways in which we were outsourcing our politics to digital media, and the implications of that for our shared future.
This book expands on the arguments of that chapter, principally Runciman’s contention that the challenges “we the people” face from AI – threats to our individuality and agency as citizens – are, while urgent, not as revolutionary as we may think. The blueprint for negotiating these challenges, Runciman believes, has been established over several centuries by the related threats from state and corporate power. The “singularity” that tech evangelists talk about – the eventual symbiosis of man and machine – would really be the “second singularity”, Runciman argues, persuasively. The first came with the age of Enlightenment, with our ability to “imagine what it would be like to organise collective enterprises as though they had the durability of machines”.
When Runciman writes and talks about politics, his year zero is frequently 1651 and the publication of Thomas Hobbes’s Leviathan, the formative meditation on the relation between institutional power and people, written in the midst of the cataclysm of civil war. Hobbes’s book gives Runciman his starting point here as well. Thinking about the supra-human digital sphere that we have created, a useful reference point, he contends, might be Hobbes’s model of the state as an “artificial man” built in the citizens’ image, but with more abstract agency. Runciman likens the idea of government to an algorithm. The Leviathan of state – or of Google or Meta – is an expression of our collective selves without a soul or a conscience. In its ideal formulation it offers continuity and shared purpose; when it goes rogue, the “man-made monster” has the capacity to exaggerate all our destructive failings.
The analogy from Hobbes allows Runciman to tease out precisely the ways that technology with a human face – the hivemind of social media, the ChatGPT robots that sound and cogitate like us, the all-seeing data-clouds of Silicon Valley – are at once similar to and distinct from those more familiar “artificial men”, the state and the corporation. Past experience teaches us how this story ends: “They are meant to work for us, but it is already possible to imagine we will end up working for them.”
Runciman’s subtitle, How We Gave Control of Our Lives to Corporations, States and AIs, makes this “handover” a fait accompli, but in the argument of his book that past tense is more provisional; the future relationship is still just about up for grabs. A lot, he argues, depends on our holding on to the semantic distinctions between decisions and knowledge, and between choices and answers. If “the computer says no” about an insurance claim, it may sound like the machine has done the deciding, but that is not quite the same as human judgment. The checks and balances we have applied to governments and corporations must be made relevant to artificial intelligence; there is a third element in those traditional relationships between the us and the Leviathans we have created.
In trying to identify what those restraints might look like, Runciman’s argument ranges far and wide, from hunter gatherers to Elon Musk, from the wisdom of juries to the (terrifying) implications of autonomous weapons systems. His conclusions about whether we will “go the way of the horse” (and be quickly obsolete beside the machines we have created) are knotty, but tend toward optimism. “If we franchise out complex decision-making to intelligent machines without abdicating personal responsibility for it, we might get the best of both worlds,” he writes – though as you read that “if”, you may equally, like me, be reminded of those drivers who have followed their satnav into a river.
• The Handover: How We Gave Control of Our Lives to Corporations, States and AIs by David Runciman is published by Profile Books (£20). To support the Guardian and Observer order your copy at guardianbookshop.com. Delivery charges may apply
