One of the most chilling moments in any sci-fi horror film is when humans and the extraterrestrials who have thus far been cloaked in shadow finally come eye to eye. As they gaze at each other, there’s that inevitable moment when they understand each character as equal and yet diabolically separate forms of intelligence (remember Sigourney Weaver and her drooling counterpart in Ridley Scott’s Alien?).
The image came to my mind as I read economist Daniel Susskind’s new book, A World Without Work: Technology, Automation and How We Should Respond, which argues that we are headed very quickly (within decades) to a time in which most human work will be displaced by machines.
In chapter four, “Understanding Machines”, Susskind retells the now fairly well-known story of AlphaGo, the computer program developed by Google’s AI subsidiary DeepMind, which beat the world champion in the ancient Chinese game of Go. It won not by playing better than the human, but by playing in a way that was inhuman.
“Almost as remarkable as its overall victory was a particular move that AlphaGo made — the 37th move in the second game — and the reaction of those watching,” writes Susskind. “Thousands of years of human play had forged a rule of thumb known even to beginners: early in the game, avoid placing stones on the fifth line from the edge. And yet, this is exactly what AlphaGo did in that move.” One expert watching called the move “beautiful”. Another said it made him feel “physically unwell”.
Those reactions encapsulate the common and diametrically opposed views of a world in which machines will do most of what human workers do today. Netscape founder and venture capitalist Marc Andreessen has famously quipped that “software is eating the world” and that in the end, there will be only two types of people left: those who program the machines, and everyone else.
Susskind goes further, arguing that we are on the verge of a new era in which the machines could come up with entirely new ideas about how to program themselves.
Susskind’s core thesis — that we are heading towards a world in which human work will become obsolete — is built on his supposition that most of the conventional notions about AI learning have been wrong.
“Economists had thought that to accomplish a task, a computer had to follow explicit rules articulated by a human being — that machine capabilities had to begin with two-down application of human intelligence.”
But Susskind believes that wisdom was false, and has led us to vastly underestimate the depth and breadth of the labour-disrupting effects of AI. According to him, many machines are “now deriving entirely new rules, unrelated to those that human beings follow. This is not a semantic quibble, but a serious shift. Machines are no longer riding on the coattails of human intelligence.”
This is, of course, the starting point of many other science fiction tales, such as Terminator. But in reality, it requires a big leap of faith. Not all technologists or economists agree that AI will be nearly as disruptive to human labour as Susskind posits (veteran venture capitalist and Cambridge university fellow Bill Janeway recently told me he thought that a large chunk of what companies call “AI” is marketing spin).
Still, there are plenty of people in the post-work world camp, including many of those at the top of the tech food chain. The parts of Susskind’s book that are most interesting and useful are those that grapple with how society should respond to that world.
Susskind makes a good case that technology is going to put Thomas Piketty’s views on steroids. Not only will the returns to capital continue to rise relative to labour, but the amount of labour itself will decrease, and eventually, disappear. Inequality, unhappiness and social unrest are the inevitable result.
We can’t educate ourselves out of the problem, at least not as we currently envision it. Neither bolstering STEM skills nor pushing liberal arts really addresses the bigger issue, which is that we need an education system that isn’t geared up to teach people how to enter the world of work, but rather how to make leisure productive. It’s a paradigm shift but, as he points out, the human race has been here before; in ancient Sparta, education was about training for war, not work. The solution most often offered by Silicon Valley — universal basic income — has possibilities, but only if tied to outcomes. As any number of interesting historical scenarios Susskind cites (1930s factory workers on benefits, Native Americans who receive government funding) prove, just giving people money doesn’t work — you have to give them meaning, too.
In Susskind’s view that will require both curbing Big Tech and, more importantly, empowering a Big State. Such a state would have to grapple with not only the challenges of redistributing a pie that, just as Keynes foretold, is big enough yet not well distributed, but also how to create meaning and assign value to citizens’ contributions in a post-work world. He summarises and expands on many of the good ideas circulating about how to do that, from creating digital sovereign wealth funds to attributing economic value to work that is currently unpaid.
There is one gaping hole in the book: China. Only three paragraphs explicitly address the fact that the Middle Kingdom is ground zero for all the changes Susskind explores — from robots replacing cheap and even higher-end labour, to the promise and perils of AI, to the ability of the state to buffer them. If there’s one place where science fiction is already reality, it is the Middle Kingdom. And the Big State is here already — it’s just headquartered in Beijing.
A World Without Work: Technology, Automation and How We Should Respond, by Daniel Susskind, Allen Lane, RRP£20, 336 pages
Get alerts on Non-Fiction when a new story is published