Celebrated English writer Ian McEwan returns with a novel pondering the ethics of artificial intelligence but set in an alternative version of 1980s Britain.
The narrator of the novel, Charlie, sees a young child being savagely whacked on the legs by his mother in a London playground for some minor disobedience. Charlie, feckless in most things, can’t help but intervene.
“‘Excuse me. Please. Please don’t do that.’ My voice sounded prissy in my ears, privileged, apologetic, lacking all authority.”
The boy’s father arrives and looks like he’s about to thump Charlie, but instead, in a surprising turn, offers up the boy in a “you’re an expert on kids, you have him” way. The proposal is quickly rescinded, though it turns out that the boy, Mark, will play a major part in Charlie’s life later on. Did that moment occur to McEwan early on in the writing process?
“No, but it’s odd how sometimes opportunities arise,” he says. “I sent my narrator for a stroll on Clapham Common to think about certain things, and once he was there, I remembered that I myself used to live near the common. I was walking by a little playground and I saw a lady smacking a young child very hard and I went and intervened. So, I just drew on my memory of that scene. Actually, there was no husband around, but she did tell me to go away in the strongest possible terms. And I just didn’t know what to do; I just felt so helpless and stupid and a middle-class person intervening in this. But she was out of control with this child, and I actually phoned Social Services afterwards.
“Having brought that scene in, I thought, aha, that little boy is going to walk into this novel in a later chapter. I hadn’t intended it up to that point, not until I had my narrator hop across the fence, as I did, and try to get between this woman and her child. There is this sort of magic circle around parents and children, and you really take your life in your hands to intervene. And it is an awkward moral question: who are you and what business is it of yours?”
We meet Charlie in a 1980s Britain reflected through a funhouse mirror. Among other changes, Margaret Thatcher has lost the Falklands War and left-winger Tony Benn is Leader of the Opposition. More crucially, Alan Turing is alive and has brought about huge advances in artificial intelligence (AI).
Charlie, in a rare moment in which he is flush with funds, buys Adam, one of the first synthetic humans. He and his sometime girlfriend, Miranda, program the ruggedly handsome Adam to be smart and, problematically, given what she will later reveal, highly ethical.
Alternative realities, particularly those that incorporate ideas utopian to the author, are often fascinating to read, but are they fun to write? “I don’t know about fun – there were pleasures attached, of course, and maybe a degree of playfulness. In some respects, it bears a relation to Nutshell [his previous book, narrated by an educated fetus], playing with the very contingent nature of reality and how things could so easily be otherwise.
“Also, having decided that I just wanted science to be in a slightly different place at a certain time, and having decided to bring Turing to life, it seemed I might as well do the same thing to politics and history itself. But I was very concerned to keep it in the background. I could so easily see that any of these events, political or military for that matter, could overwhelm this novel and knock me off my course. Which was what it would be to have a close relationship to an entirely artificial human who was completely plausible.”
McEwan did toy with other eventualities. “Of course, I was speculating like crazy while lying in the bath, but once it got to the keyboard or the pen, things solidified and all the other ghosts melted away.”
“I’m interested in a human project of the imagination that’s been going on for maybe 3000 years. We have the myth of Prometheus, fashioning an artificial human; the bronze automaton Talos from Greek mythology; and the Genesis story of Adam and Eve being fashioned out of clay.
“So, long before Mary Shelley’s Frankenstein, there’s been this very strong idea that we could make ourselves and we could become as God. And so, when the technological ability is finally there to make a completely plausible, intelligent, emotionally warm and convincing human being, and we have to interact with it, we will be enacting what we’re already just beginning to do.”
Which is to interface with machines. “Talking to them down the phone, wondering who should live or die in a car accident, and what we’re going to tell autonomous vehicles to do. A world in which these machines are running our electricity grids and our sewage farms and flying our aeroplanes. I think it’s going to consume us for two or three centuries, because a plausible human being is a long way off yet.
“When you think of the processing power of the human brain, with its hundred billion neurons and an average of 7000 connections per neuron, all running on 25 watts – the strength of a very dim light bulb that never overheats – we’re nowhere near it. But we are already in the arena. We’re already down there with the machines, and it’s going to be, as well as a technological problem, much more interestingly, an ethical one.
“I’ve tried to jump ahead and say, well, what happens when we have an old-fashioned love triangle? What is the conversation like between the two humans when one of them has had sex with an artificial human? Is it a betrayal or not? That’s my starting point, really, to get in as close as I could, as close as a novel can, to all the fine-grained relationships we have with each other, to see what it would be like if they were exchanged with a machine.”
Is he hopeful or fearful of AI? Will it be capable of self-harm, or falling in love, or sorrow?
“I’m interested in the question. I just don’t know – I’m really a materialist in these things. If you have a machine that could process like a human brain can, then you would have all the things a human brain can do, which is the mind. So, in a sense, it’s a sort of technical matter: could you make a one-litre computer? Visual processing takes up a third of our neural capacities – even that is way beyond what a machine could do.
“To run something like our processing power, you’d probably need something like a room full of computers, generating an enormous amount of heat. My Adam can go for a 17km run in two hours, or talk non-stop for 12 days. We don’t even have a battery [suitable]; we just have these sad lithium-ion things that weigh a lot.”
But we’re already taking baby steps, he says. They could be difficult steps, such as the two recent aeroplane crashes that appear to have been caused by the plane’s “brain” telling the pilots it was stalling, or Hal in Stanley Kubrick’s 2001, which decides your fate and it’s not to your liking. “But, they could be wonderful. All our philosophies and religions tell us how to be good, but we’re not very good at being good. We could imagine programming an artificial human with all our best moral precepts. We might get rather annoyed at it when it starts telling us that we’re not behaving as well as we should.”
Early on, we are reminded about Isaac Asimov’s first law of robotics, which says, “A robot may not injure a human being or, through inaction, allow a human being to come to harm.” And yet, Adam breaks Charlie’s wrist.
“Unintentionally,” says McEwan. “He doesn’t mean to. The narrator decides to turn Adam off for a bit. But Adam has come to that point in his learning curve when he has concluded that consciousness is his greatest gift and he’s not going to let anybody have control over it, other than himself. And so he resists Charlie, the narrator, and breaks his wrist.”
What’s odd is that Charlie seems to take the act quite lightly. “Well, I broke my wrist as a child,” says McEwan. “Once your arm is in a cast, life just goes on, I guess.”
Still, a sociopathic robot is a terrifying prospect, he agrees. “We could send cruel, unfeeling artificial humans into battle. It’s a double-edged sword.”
Despite its speculative premise, McEwan doesn’t think of the novel as science fiction. “I don’t have much interest in science fiction, really. It doesn’t seem to me human enough. I’ve relied on many of the techniques of the novel: the stranger who appears in your midst; the three-cornered love affair; the growth of emotional relationships, and so on. Science fiction, by putting things in the future, tends to have a rather predictive quality, and I know we’ve been talking here in predictive terms, but I really wanted to concentrate not on ‘this may or may not happen’ but just to enter into what it would be like to treat Adam as human. I tried to make him as human as possible. I used every last effort as a novelist to make the reader suffer the same dilemmas as the narrator. Half the time you’re thinking, ‘He’s just a machine, I’m just playing a computer game.’ And then being caught short, saying, ‘I can’t help treating him like a human and I’m coming to really like him, even love him.’”
The narrators of McEwan’s novels often come across as eminently reasonable and sophisticated. Is he at all like Adam, the most rational voice of the novel? “You’ve got to remember that there are four characters in this novel: Adam; the child; the narrator, who also has real emotional problems; and a young woman. And I’m part of all of them.”
Machines Like Me, by Ian McEwan (Penguin Random House, $48), is out now.
This article was first published in the April 20, 2019 issue of the New Zealand Listener.