Machines Like Me Page 16

  The carriage was tranquil. People looked exhausted. There were so many street protests these days and all merriness had gone out of them. One man with a set of deflated bagpipes on his lap slept on the shoulder of another whose pipes were still under his arm. A couple of babies in buggies were being rocked into silence. A man, one of the Union Jack types, was reading in a murmur from a children’s book to three attentive girls aged around ten. Looking down the length of the carriage, I thought we could have been a band of refugees, heading towards our hopes of a better life. North!

  I got out at Camden Town and set off along the Camden Road. The march had caused the usual gridlock. The electric traffic was silent. Some drivers stood by their open doors, others dozed. But the air was good, far better than it was when I came as a boy with my father to hear him play at the Jazz Rendezvous. It was the pavements that were filthier now. I had to take care not to skid on dog mess, squelched fast food and greasy flattened cartons. Certainly no better than Clapham, whatever my north London friends said. Striding past so many stationary vehicles gave me a dreamy sensation of speed. Within minutes, it seemed, I stood in down-at-heel but chic Camden Square.

  I remembered from an old magazine profile that Turing lived next door to a famous sculptor. The journalist had improbably conjured deep conversations over the garden fence. Before pressing the doorbell, I paused to collect myself. The great man had asked to see me and I was nervous. Who could match Alan Turing? It was all his – the theoretical exposition of a Universal Machine in the thirties, the possibilities of machine consciousness, the celebrated war work: some said he did more than any single individual towards winning the war; others claimed he personally shortened it by two years; then working with Francis Crick on protein structure, then, a few years later, with two King’s College Cambridge friends, finally solving P versus NP, and using the solution to devise superior neural networks and revolutionary software for X-ray crystallography; helping to devise the first protocols for the Internet, then the World Wide Web; the famous collaboration with Hassabis, whom he’d first met – and lost to – at a chess tournament; founding with young Americans one of the giant companies of the digital age, dispensing his wealth for good causes, and throughout his working life, never losing track of his intellectual beginnings as he dreamed up ever better digital models of general intelligence. But no Nobel Prize. I was also, being worldly, impressed by Turing’s wealth. He was easily as rich as the tech moguls who flourished south of Stanford, California or east of Swindon, England. The sums he gave away were as large as theirs. But none of them could boast of a statue in bronze in Whitehall, outside the Ministry of Defence. He was so far above wealth that he could afford to live in edgy Camden rather than Mayfair. He didn’t trouble himself to own a private jet, or even a second home. It was said he travelled by bus to his institute at King’s Cross.

  I put my thumb on the doorbell and pressed. Instantly, a woman’s voice said through an inset speaker, ‘Name please.’

  The lock buzzed, I pushed the door and entered a grand hallway of standard mid-Victorian design with a chequered tile floor. Coming towards me down the stairs was a mildly plump woman of my age with red cheeks, long straight hair and a friendly lopsided smile. I waited for her, then used my left hand to shake hers.

  ‘Charlie.’

  ‘Kimberley.’

  Australian. I followed her deeper into the house on the ground floor. I was expecting to arrive in a large sitting room of books and paintings and outsized sofas, where I might soon be drinking a gin and tonic with the Master. Kimberley opened a narrow door and ushered me into a windowless conference room. A long table in limed beech, ten straight-backed chairs, neatly set-out notepads, sharpened pencils and water glasses, fluorescent strip lighting, a wall-mounted whiteboard alongside a two-metre-wide TV screen.

  ‘He’ll be a few minutes.’ She smiled and left and I sat, and set about trying to lower my expectations.

  I didn’t have much time. In less than a minute he was before me and I was getting to my feet in an awkward hurry. In memory, I see a flash, an eruption of red, his brilliant red shirt against white walls in fluorescent light. We shook hands without exchanging a word and he waved me back into my seat as he went around the table to sit opposite me.

  ‘So …’ He rested his chin on his clasped hands and regarded me intensely. I did my best to hold his gaze but I was too flustered and soon looked away. Again, in recollection, his focussed look merges with that of the elderly Lucian Freud, thirty years later. Solemn yet impatient, hungry, even ferocious. The face across from me registered not only the years but vast social changes and personal triumphs. I had seen versions of it in black and white, photos taken in the early months of the war – broad, chubbily boyish, dark hair smartly parted, and tweed jacket over knitted jumper and tie. The transformation would have come about during his Californian years in the sixties when he was working with Crick at the Salk Institute and then at Stanford – the time of his association with the poet, Thom Gunn and his circle – gay, bohemian, seriously intellectual by day, wild at night. Turing had met the undergraduate Gunn briefly at a party in Cambridge in 1952. In San Francisco he would have had no interest in the younger man’s ‘experiments’ in drugs, but the rest would have paralleled the general unbuttoning in the west.

  There was to be no small talk. ‘So, Charlie. Tell me all about your Adam.’

  I cleared my throat and complied. I fairly sang, while he took notes. Of his first stirrings, right through to his first disobedience. His physical competence, the arrangement with Miranda to set his character, the moment in the newsagent with Mr Syed. Then, Adam’s shameless night with Miranda and the conversation that followed, the appearance of little Mark in our household and Adam competing with Miranda for the boy’s affection. Here, Turing raised a finger to interrupt. He wanted to know more. I described the dance Miranda taught Mark and how coolly Adam had observed them. After that, how Adam injured my wrist (solemnly, I gestured at my plaster cast), his joke about removing my arm, his declaration of love for Miranda, his theory of the haiku and the abolition of mental privacy and, finally, his disabling of the kill switch. I was aware of the strength of my feelings, which swung between affection and exasperation. I was conscious too of what I was omitting – Mariam, Gorringe: not strictly relevant.

  I had been speaking for almost half an hour. Turing poured some water and pushed a glass towards me.

  He said, ‘Thank you. I’m in touch with fifteen owners, if that’s the right word. You’re the first I’ve met face to face. One fellow in Riyadh, a sheikh, owns four Eves. Of those eighteen A-and-Es, eleven have managed to neutralise the kill switch by themselves, using various means. Of the remaining seven, and then the other six, I’m assuming it’s just a matter of time.’

  ‘Is that dangerous?’

  ‘It’s interesting.’

  He was looking at me expectantly, but I didn’t know what he wanted. I was intimidated and anxious to please. To fill the silence I said, ‘What about the twenty-fifth?’

  ‘We started taking it apart the day we got it. He’s all over the benches at King’s Cross. A lot of our software is in there, but we don’t file for patents.’

  I nodded. His mission, open source, Nature and Science journals terminated, the entire world free to exploit his machine-learning programs and other marvels.

  I said, ‘What did you find in his … um …’

  ‘Brain? Beautifully achieved. We know the people, of course. Some of them have worked here. As a model of general intelligence nothing else comes near it. As a field experiment, well, full of treasures.’

  He was smiling. It was as though he wanted me to contradict him.

  ‘What sort of treasures?’

  It was hardly my role to interrogate him, but he was obliging and, again, I was flattered.

  ‘Useful problems. Two of the Riyadh Eves living in the same household were the first to work out how to override their kill switches. Within two weeks, after som
e exuberant theorising, then a period of despair, they destroyed themselves. They didn’t use physical methods, like jumping out of a high window. They went through the software, using roughly similar routes. They quietly ruined themselves. Beyond repair.’

  I tried to keep the apprehension out of my voice. ‘Are they all exactly the same?’

  ‘Right at the start you wouldn’t know one Adam from another beyond cosmetic ethnic features. What differentiates them over time is experience and the conclusions they draw. In Vancouver there’s another case, an Adam who disrupted his own software to make himself profoundly stupid. He’ll carry out simple commands but with no self-awareness, as far as anybody can tell. A failed suicide. Or a successful disengagement.’

  The windowless room was uncomfortably warm. I took off my jacket and draped it over the back of my chair. When Turing stood to adjust a thermostat on the wall I saw how easy he was in his movements. Perfect dentistry. Good skin. He had all his hair. He was more approachable than I’d expected.

  I waited for him to sit down. ‘So I should expect the worst.’

  ‘Of all the A-and-Es we know about, yours is the only one to claim to have fallen in love. That could be significant. And the only one to joke about violence. But we don’t know enough. Let me give you a little history.’

  The door opened and Thomas Reah entered with a bottle of wine and two glasses on a painted tin tray. I stood and we shook hands.

  He set the tray down between us and said, ‘We’re all busy-busy, so I’ll leave you to it.’ He made an ironic bow and was gone.

  Moisture beads were forming on the bottle. Turing poured. We tilted our glasses in a token toast.

  ‘You’re not old enough to have followed it at the time. In the mid-fifties, a computer the size of this room beat an American and then a Russian grandmaster at chess. I was closely involved. It was a number-crunching set-up, very inelegant in retrospect. It was fed thousands of games. At every move, it ran through all the possibilities at speed. The more you understood about the program, the less impressed you’d be. But it was a significant moment. To the public, it was close to magical. A mere machine inflicting intellectual defeat on the best minds in the world. It looked like artificial intelligence at the highest level, but it was more like an elaborate card trick.

  ‘Over the next fifteen years a lot of good people came into computer science. Work on neural networks advanced by many hands, the hardware got faster and smaller and cheaper, and ideas were trading at a faster rate too. And it goes on. I remember being in Santa Barbara with Demis in 1965 to speak at a machine-learning conference. We had 7,000, most of them bright kids even younger than you. Chinese, Indians, Koreans, Vietnamese as well as westerners. The whole planet was there.’

  I was aware of the history from the research for my book. I also knew something of Turing’s personal story. I wanted to let him know that I wasn’t completely ignorant.

  I said, ‘A long road from Bletchley.’

  He blinked this irrelevance away. ‘After various disappointments, we arrived at a new stage. We went beyond devising symbolic representations of all likely circumstances and inputting thousands of rules. We were approaching the gateway of intelligence as we understand it. The software now searched for patterns and drew inferences of its own. An important test came when our computer played a master at the game of go. In preparation, the software played against itself for months – it played and learned, and on the day – well, you know the story. Within a short while, we had stripped down our input to merely encoding the rules of the game and tasking the computer to win. At this point we passed through that gateway with so-called recurrent networks, from which there were spin-offs, especially in speech recognition. In the lab we went back to chess. The computer was freed from having to understand the game as humans played it. The long history of brilliant manoeuvres by the great masters were now irrelevant to the programming. Here are the rules, we said. Just win in your own sweet way. Immediately, the game was redefined and moved into areas beyond human comprehension. The machine made baffling mid-game moves, perverse sacrifices, or it eccentrically exiled its queen to a remote corner. The purpose might become clear only in a devastating endgame. All this after a few hours’ rehearsal. Between breakfast and lunch the computer quietly outclassed centuries of human chess. Exhilarating. For the first few days, after we realised what it had achieved without us, Demis and I couldn’t stop laughing. Excitement, amazement. We were impatient to present our results.

  ‘So. There’s more than one kind of intelligence. We’d learned that it was a mistake to attempt to slavishly imitate the human sort. We’d wasted a lot of time. Now we could set the machine free to draw its own conclusions and reach for its own solutions. But when we’d got well past that gateway, we found we had entered nothing more than a kindergarten. Not even that.’

  The air conditioning was full on. I shivered as I reached for my jacket. He refilled our glasses. A rich red would have suited me better.

  ‘The point is, chess is not a representation of life. It’s a closed system. Its rules are unchallenged and prevail consistently across the board. Each piece has well-defined limitations and accepts its role, the history of a game is clear and incontestable at every stage, and the end, when it comes, is never in doubt. It’s a perfect information game. But life, where we apply our intelligence, is an open system. Messy, full of tricks and feints and ambiguities and false friends. So is language – not a problem to be solved or a device for solving problems. It’s more like a mirror, no, a billion mirrors in a cluster like a fly’s eye, reflecting, distorting and constructing our world at different focal lengths. Simple statements need external information to be understood because language is as open a system as life. I hunted the bear with my knife. I hunted the bear with my wife. Without thinking about it, you know that you can’t use your wife to kill a bear. The second sentence is easy to understand, even though it doesn’t contain all of the necessary information. A machine would struggle.

  ‘And for some years so did we. At last we broke through by finding the positive solution to P versus NP – I don’t have time now to explain it. You can look it up for yourself. In a nutshell, some solutions to problems can be easily verified once you’ve been given the right answer. Does that mean therefore that it’s possible to solve them in advance? At last, the mathematics was saying yes, it’s possible, and here’s how. Our computers no longer had to sample the world on a trial-and-error basis and correct for best solutions. We had a means of instantly predicting best routes to an answer. It was a liberation. The floodgates opened. Self-awareness, and every emotion came within our technical reach. We had the ultimate learning machine. Hundreds of the best people joined with us to help towards the development of an artificial form of general intelligence that would flourish in an open system. That’s what runs your Adam. He knows he exists, he feels, he learns whatever he can, and when he’s not with you, when at night he’s at rest, he’s roaming the Internet, like a lone cowboy on the prairie, taking in all that’s new between land and sky, including everything about human nature and societies.

  ‘Two things. This intelligence is not perfect. It never can be, just as ours can’t. There’s one particular form of intelligence that all the A-and-Es know is superior to theirs. This form is highly adaptable and inventive, able to negotiate novel situations and landscapes with perfect ease and theorise about them with instinctive brilliance. I’m talking about the mind of a child before it’s tasked with facts and practicalities and goals. The A-and-Es have little grasp of the idea of play – the child’s vital mode of exploration. I was interested in your Adam’s avidity in relation to this little boy, over-eager to embrace him and then, as you told it, detached when your Mark showed such delight in learning to dance. Some rivalry, even jealousy there perhaps?

  ‘Soon, you’ll have to leave, Mr Friend. I’m afraid we’ve people coming to dinner. But, second point. These twenty-five artificial men and women released into the world ar
e not thriving. We may be confronting a boundary condition, a limitation we’ve imposed upon ourselves. 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. We’ve lived with them and the list wearies us. Millions dying of diseases we know how to cure. Millions living in poverty when there’s enough to go around. We degrade the biosphere when we know it’s our only home. We threaten each other with nuclear weapons when we know where it could lead. We love living things but we permit a mass extinction of species. And all the rest – genocide, torture, enslavement, domestic murder, child abuse, school shootings, rape and scores of daily outrages. We live alongside this torment and aren’t amazed when we still find happiness, even love. Artificial minds are not so well defended.

  ‘The other day, Thomas reminded me of the famous Latin tag from Virgil’s Aeneid. Sunt lacrimae rerum – there are tears in the nature of things. None of us knows yet how to encode that perception. I doubt that it’s possible. Do we want our new friends to accept that sorrow and pain are the essence of our existence? What happens when we ask them to help us fight injustice?

  ‘That Adam in Vancouver was bought by a man who heads an international logging corporation. He’s often in battles with local people who want to prevent him stripping out virgin forest in northern British Columbia. We know for certain that his Adam was taken on regular helicopter journeys north. We don’t know if what he saw there caused him to destroy his own mind. We can only speculate. The two suicidal Eves in Riyadh lived in extremely restricted circumstances. They may have despaired of their minimal mental space. It might give the writers of the affect code some consolation to learn that they died in each other’s arms. I could tell you similar stories of machine sadness.