August 2019

Arts & Letters

Robot love: Ian McEwan’s ‘Machines Like Me’ and Jeanette Winterson’s ‘Frankissstein’

By Karen Hitchcock
Literary authors tackle sentience and rationality in AI, with horrific results

In the past if you wanted to read a story peppered with robot characters you’d head for the science-fiction section. But somewhere along the line apocalypse-threatening robots started invading literary fiction. In February this year, The New Yorker published a story by T. Coraghessan Boyle, “Asleep at the Wheel”, featuring an AI-powered car named Carly that starts dictating how its owner should behave, by, for example, locking her in and out of the car to prevent her consorting with a man it doesn’t approve of. Then, Jeanette Winterson and Ian McEwan both released books on the subject, Frankissstein: A Love Story and Machines Like Me, respectively. It’s in the air. The possibility of sentient robots is no longer as speculative.

The prospect of a true, general artificial intelligence – that is, a machine with the capacity to think, learn and reason flexibly, and with sentience – was never one that worried me much. I had vague notions that consciousness was an irreplicable mystery, and that any intelligent computer could simply be programmed to limit its scope and serve our whims. Then scientists and “thought leaders” – Elon Musk, Sam Harris, Bill Gates, Nick Bostrom and even Stephen Hawking – started speaking of the technology as if it were imminent and would probably mark the end of our time at the top of the food chain, and turn us into what Musk described as “the biological boot loader for a digital superintelligence”.

Limited AI is already everywhere – used for targeted advertising, news curation, car safety programs and to power search engines. The kind of AI that has smart people nervous is one with the capacity to self-improve rapidly and develop autonomously. And right now there is an unregulated global race among gigantic tech companies to produce such a machine – fuelled by the thrill of invention and the motivations of commerce. We don’t really know what’s going on at DeepMind, a company owned by Google, or in other laboratories at the cutting edge of AI advancement. Those labs are not required to follow specific guidelines or safety standards, they have no global oversight, closet their advances and have been slow to engage with ethics departments. Should we be concerned?

Thinking about intelligent robots, trying to build them, demands a precise understanding of what it is to be human. It’s useful to know how it is that we come to think and learn and experience our self and the world in order to build something with the capacity to do so. Can a robot learn to love? Well, what exactly is love? AI researchers are bringing in developmental psychologists and philosophers to collaborate on how, for example, curiosity, play and open-system learning work in us. And thinking through the consequences to humanity if we ever do manage to create such machines requires big imaginations.

Both McEwan’s and Winterson’s books feature cameos from key historical figures from the history of computer science (Turing, Musk and Kurzweil all have roles). The narratives are both driven by human–human and human–robot love stories, have dystopic outcomes and, when read in conjunction, cover most contemporary thinking about AI. They are both enormously clever,carefully researched (and enjoyable) novels. This grappling with what it is to be human, especially in the context of AI, is ideal material for the novelist.

Winterson’s book has two narrative threads that play out in alternating chapters. The first is set in 1816 and is loosely based around the actual events surrounding Mary Shelley’s writing of Frankenstein. She, her husband (the poet Percy Shelley), her sister, Lord Byron and his personal physician are holed up in a mansion during a weeks-long storm. For entertainment, Byron sets them the task of writing a ghost story. Frankenstein is Mary Shelley’s contribution. Winterson alternates the Shelley chapters with a story set in the modern world: a wry, non-binary narrator, Ry Shelley, is writing an article about the development of robots. It’s a brilliant juxtaposition. In the 1800s galvanism promised to animate the corpse through the application of electrical currents. But a question remained, one that Frankenstein can be thought to muse upon: would such a creature have a soul? In the contemporary world we seek to create digital minds and are faced with the same question, albeit in a secular cloak: would such a creature have consciousness?

Trying to answer this question of the possibility of intelligent and potentially conscious machines sends one down a spiralling abstract hole, where assumptions are smashed, bearings are lost and everything is contestable. Are our brains just mega-powerful meat-computers or is intelligence more than just the processing of information? How will we know if a machine is truly thinking or merely “acting” as if it is thinking? What is it to “truly think”? Is consciousness just the outcome of many neurons firing? Is it necessary for intelligence? We don’t know how consciousness emerges from the material world but we know from our personal internal experience that it does. I know that I am conscious. But what proof do I have that even you, an actual human being, are conscious, let alone know it of a robot? Winterson and McEwan each ground these questions in a story, and gently lead the reader through the philosophy and science of intelligent machines as the characters enact the conclusions. You’re barely aware that you’re being walked through a series of contemporary bioethical dilemmas.

The Ry chapters of Winterson’s book take place in a world where only simple robots are widely available (mainly female sexbots aimed at the straight male market) and a true AI has not yet been created. In Ry’s world as much scientific effort is being placed upon ways to ensure immortality as it is upon developing an AI: for example, through cryogenic preservation of the dead for future resuscitation and the “uploading” of minds onto digital platforms. In one of many nods to Frankenstein, the cutting edge of this work is led by a renegade scientist named Victor. The novel approximates the moment we now find ourselves in. It is highly conscious of gender, and science is repeatedly framed as being driven by male fantasies: to usurp women’s role in the creation of life, to create perfect artificial sex partners. (There is no mention in the book of perfect sex partners for the female, though I imagine there would be an equally vigorous market for something reliable.) In the world of Winterson’s book, if we create a mind from a computer it follows that we can shift our selves into the digital realm, a limitless universe of pure consciousness, untethered from biological limitations, free from “these caravans of tissue and bone”. Free from labels, appearance and from sensation as we now know it.

McEwan sets his book in an alternative 1980s Thatcher-era England where functioning robots already exist and have replaced many manual labour jobs, resulting in violent riots by disenfranchised workers. Things are not going well in society when the first, small batch of humanoid robots are released for sale. “Global temperatures rose … Everything was rising – hopes and despair, misery, boredom and opportunity. There was more of everything. It was a time of plenty.” The protagonist, Charlie, spends his entire inheritance to secure one – an Adam, though he would have preferred one of the sold-out Eves. Adam can pass for human. It is warm, it breathes, it quickly masters social conventions, writes poetry, feels sensation and emotion, and it can fuck. Adam’s position in the house moves from child to maid to autonomously acting agent. Adam has an off button, which it disables early in the piece (as do many of the other 24 robots in this first release). Imagine trying to curb the enthusiasms of your teenage child if she had extraordinary brain power, unlimited physical strength and was not financially dependent on you. There is much havoc to be wreaked.

No buyer expected the Adam and Eve robots to become sentient, to have the capacity to suffer, desire and reflect. When Adam displays behaviours that imply sentience, Charlie tells himself that these are merely the result of sophisticated programming rather than a feeling entity. He unsuccessfully battles with his spontaneous emotional responses to the machine. Reams of behavioural research show that even if robots are not conscious, or do not possess generalised intelligence, we will respond to them as if they are and do. People already yell at their remote controls.

After his girlfriend has sex with Adam (Charlie listens to her shrieks of pleasure from the next room and becomes enraged), she says, “If I’d gone to bed with a vibrator would you be feeling the same? … He has as much consciousness as one.” It’s not the same, but Charlie cannot articulate why that is so without granting Adam the status of subject. He meets with Alan Turing, who has been closely monitoring the Adams’ and Eves’ induction to society, and it is from Turing that Charlie hears the fate of the other robots. Turing says:

These twenty-five artificial men and women released into the world are not thriving … We create a machine with intelligence and self-awareness and push it out into our imperfect world. Devised along generally rational lines, well disposed to others, such a mind soon finds itself in a hurricane of contradictions … Millions dying of diseases we know how to cure. Millions living in poverty when there’s enough to go round … there’s nothing in all their beautiful code that could prepare Adam and Eve for Auschwitz.

For McEwan, AI – devoid of all the wily defences that our minds draw upon to protect us – reflects back to us the true horrors that we are. For Winterson, the robot is a monster-child built out of and built to serve male desires.

Intellectuals from a variety of disciplines tie themselves in knots trying to work out the signature of consciousness, to devise tests to prove it is present. It is the single thing we can each be sure of, and yet it eludes objective proof. How will we know if the artificial thinkers we produce have the capacity to truly feel? In the hands of these novelists – Shelley, Winterson and McEwan – the answer is clear: we can be sure of a machine’s sentience when it seeks its own destruction. Turing tells Charlie that the robots have started to suicide by purposefully destroying their programs. These perfect minds are too perceptive and rational to destroy the world. And too perceptive and rational to go on living in it. A sexbot in Frankissstein feels sad when her owner dies and she’s sold on eBay to a “fuck only type”. She wishes she could wipe her software.

Of course, Frankenstein’s monster also intends to kill himself in despair at the end of Shelley’s novel. Machines Like Me and Frankissstein both owe huge debts to Mary Shelley’s creation. We are still galvanists, trying to spark life out of wire and electricity, unsure of what we will do should we be successful. McEwan and Winterson have just updated the language of the warning: careful what you wish for.

Karen Hitchcock

Karen Hitchcock is a doctor and writer. She is the author of a collection of short fiction, Little White Slips, and the Quarterly Essay Dear Life: On Caring for the Elderly.

Image of The Monthly cover August 2019

August 2019

From the front page

Hard-pressed

The government appears to be dragging its heels on media law reform

Photograph of Harold Bloom

Canon salute

Remembering Harold Bloom (July 11, 1930 – October 14, 2019)

The NBN-ding story

New developments in the interminable debate over broadband in Australia

‘The weekend’ cover

‘The Weekend’ by Charlotte Wood

The Stella Prize–winner returns with a stylish character study of women surprised by age


In This Issue

Illustration

The terrible truth of climate change

The latest science is alarming, even for climate scientists

Photo of Lil Nas X

Happy trails: Lil Nas X’s ‘Old Town Road’

The gay country-rapper exposes the complex play of identity, algorithms and capitalism

Illustration

Surf and turfed

The Australian surfers battling Chinese developers in Fiji

Cover of ‘The White Girl’ by Tony Birch

‘The White Girl’ by Tony Birch

An emotionally eloquent novel about the necessary inheritance of strength in Indigenous women


More in Arts & Letters

Still from Todd Phillips’ ‘Joker’

No one’s laughing now: Todd Phillips’ ‘Joker’

A gripping psychological study of psychosis offers a surprising change of pace in the superhero genre

‘Penny Wong: Passion and Principle’

‘Penny Wong: Passion and Principle’

Margaret Simons’ biography of one of the country’s most admired politicians

Patricia Cornelius

Patricia Cornelius: No going gently

‘Anthem’ marks the return of the Australian playwright’s working-class theatre

East Melbourne liturgy


More in Books

‘Penny Wong: Passion and Principle’

‘Penny Wong: Passion and Principle’

Margaret Simons’ biography of one of the country’s most admired politicians

Image of ‘Sex in the Brain’

Our largest sexual organ: Amee Baird’s ‘Sex in the Brain’

We know surprisingly little about how our brains orchestrate our sex lives

Cover image of Underland by Robert Macfarlane

The chthonic realms explored in Robert Macfarlane’s ‘Underland’

Cave systems, mines, urban sewers, mycelial networks, moulins and more

Book cover of Choice Words

The desperate, secretive drama: ‘Choice Words’ edited by Louise Swinn

Personal stories consider questions of choice, legality and stigma surrounding abortion


Read on

Photograph of Harold Bloom

Canon salute

Remembering Harold Bloom (July 11, 1930 – October 14, 2019)

Image from ‘Judy’

Clang, clang, clang: ‘Judy’

The Judy Garland biopic confuses humiliation for homage

Image of Joel Fitzgibbon and Anthony Albanese

Climate of blame

Labor runs the risk of putting expediency over principle

Afterwards, nothing is the same: Shirley Hazzard

On the splendour of the acclaimed author’s distinctly antipodean seeing


×
×