Hello World Read online

  About the Book

  Would you let your family’s full medical history be made public if it would help find a cure for cancer?

  You are accused of a crime. Who do you want to decide your future – a human or a machine?

  Which driverless car would you buy – one that prioritizes your life in the event of a crash, or one that reacts to save as many lives as possible?

  Welcome to the age of the machine, the story of a not-too-distant future where algorithms rule supreme, making decisions about our healthcare, security, what we watch, where we go, even who goes to prison. So how much should we rely on them? What kind of future do we want?

  Hannah Fry takes us on a tour of the good, the bad and the downright ugly of the algorithms that surround us. In Hello World she lifts the lid on their inner workings, demonstrates their power, exposes their limitations, and examines whether they really are an improvement on the humans they are replacing.

  Contents

  Cover

  About the Book

  Title Page

  Dedication

  A note on the title

  Introduction

  Power

  Data

  Justice

  Medicine

  Cars

  Crime

  Art

  Conclusion

  Photograph Credits

  Notes

  Acknowledgements

  Index

  About the Author

  Also by Hannah Fry

  Copyright

  For Marie Fry.

  Thank you for never taking no for an answer.

  A note on the title

  WHEN I WAS 7 years old, my dad brought a gift home for me and my sisters. It was a ZX Spectrum, a little 8-bit computer – the first time we’d ever had one of our own. It was probably already five years out of date by the time it arrived in our house, but even though it was second-hand, I instantly thought there was something marvellous about that dinky machine. The Spectrum was roughly equivalent to a Commodore 64 (although only the really posh kids in the neighbourhood had one of those) but I always thought it was a far more beautiful beast. The sleek black plastic casing could fit in your hands, and there was something rather friendly about the grey rubber keys and rainbow stripe running diagonally across one corner.

  For me, the arrival of that ZX Spectrum marked the beginning of a memorable summer spent up in the loft with my elder sister, programming hangman puzzles for each other, or drawing simple shapes through code. All that ‘advanced’ stuff came later, though. First we had to master the basics.

  Looking back, I don’t exactly remember the moment I wrote my first ever computer program, but I’m pretty sure I know what it was. It would have been the same simple program that I’ve gone on to teach all of my students at University College London; the same as you’ll find on the first page of practically any introductory computer science textbook. Because there is a tradition among all those who have ever learned to code – a rite of passage, almost. Your first task as a rookie is to program the computer to flash up a famous phrase on to the screen:

  ‘HELLO WORLD’

  It’s a tradition that dates back to the 1970s, when Brian Kernighan included it as a tutorial in his phenomenally popular programming textbook.1 The book – and hence the phrase – marked an important point in the history of computers. The microprocessor had just arrived on the scene, heralding the transition of computers from what they had been in the past – enormous great specialist machines, fed on punch cards and ticker-tape – to something more like the personal computers we’re used to, with a screen, a keyboard and a blinking cursor. ‘Hello world’ came along at the first moment when chit-chat with your computer was a possibility.

  Years later, Brian Kernighan told a Forbes interviewer about his inspiration for the phrase. He’d seen a cartoon showing an egg and a newly hatched chick chirping the words ‘Hello world!’ as it was born, and it had stuck in his mind.

  It’s not entirely clear who the chick is supposed to be in that scenario: the fresh-faced human triumphantly announcing their brave arrival to the world of programming? Or the computer itself, awakening from the mundane slumber of spreadsheets and text documents, ready to connect its mind to the real world and do its new master’s bidding? Maybe both. But it’s certainly a phrase that unites all programmers, and connects them to every machine that’s ever been programmed.

  There’s something else I like about the phrase – something that has never been more relevant or more important than it is now. As computer algorithms increasingly control and decide our future, ‘Hello world’ is a reminder of a moment of dialogue between human and machine. Of an instant where the boundary between controller and controlled is virtually imperceptible. It marks the start of a partnership – a shared journey of possibilities, where one cannot exist without the other.

  In the age of the machine, that’s a sentiment worth bearing in mind.

  Introduction

  ANYONE WHO HAS ever visited Jones Beach on Long Island, New York, will have driven under a series of bridges on their way to the ocean. These bridges, primarily built to filter people on and off the highway, have an unusual feature. As they gently arc over the traffic, they hang extraordinarily low, sometimes leaving as little as 9 feet of clearance from the tarmac.

  There’s a reason for this strange design. In the 1920s, Robert Moses, a powerful New York urban planner, was keen to keep his newly finished, award-winning state park at Jones Beach the preserve of white and wealthy Americans. Knowing that his preferred clientele would travel to the beach in their private cars, while people from poor black neighbourhoods would get there by bus, he deliberately tried to limit access by building hundreds of low-lying bridges along the highway. Too low for the 12-foot buses to pass under.1

  Racist bridges aren’t the only inanimate objects that have had a quiet, clandestine control over people. History is littered with examples of objects and inventions with a power beyond their professed purpose.2 Sometimes it’s deliberately and maliciously factored into their design, but at other times it’s a result of thoughtless omissions: just think of the lack of wheelchair access in some urban areas. Sometimes it’s an unintended consequence, like the mechanized weaving machines of the nineteenth century. They were designed to make it easier to create complicated textiles, but in the end, the impact they had on wages, unemployment and working conditions made them arguably more tyrannical than any Victorian capitalist.

  Modern inventions are no different. Just ask the residents of Scunthorpe in the north of England, who were blocked from opening AOL accounts after the internet giant created a new profanity filter that objected to the name of their town.3 Or Chukwuemeka Afigbo, the Nigerian man who discovered an automatic hand-soap dispenser that released soap perfectly whenever his white friend placed their hand under the machine, but refused to acknowledge his darker skin.4 Or Mark Zuckerberg, who, when writing the code for Facebook in his dorm room in Harvard in 2004, would never have imagined his creation would go on to be accused of helping manipulate votes in elections around the globe.5

  Behind each of these inventions is an algorithm. The invisible pieces of code that form the gears and cogs of the modern machine age, algorithms have given the world everything from social media feeds to search engines, satellite navigation to music recommendation systems, and are as much a part of our modern infrastructure as bridges, buildings and factories ever were. They’re inside our hospitals, our courtrooms and our cars. They’re used by police forces, supermarkets and film studios. They have learned our likes and dislikes; they tell us what to watch, what to read and who to date. And all the while, they have the hidden power to slowly and subtly change the rules about what it means to be human.

  In th
is book, we’ll discover the vast array of algorithms on which we increasingly, but perhaps unknowingly, rely. We’ll pay close attention to their claims, examine their undeclared power and confront the unanswered questions they raise. We’ll encounter algorithms used by police to decide who should be arrested, which make us choose between protecting the victims of crime and the innocence of the accused. We’ll meet algorithms used by judges to decide on the sentences of convicted criminals, which ask us to decide what our justice system should look like. We’ll find algorithms used by doctors to over-rule their own diagnoses; algorithms within driverless cars that insist we define our morality; algorithms that are weighing in on our expressions of emotion; and algorithms with the power to undermine our democracies.

  I’m not arguing that algorithms are inherently bad. As you’ll see in these pages, there are many reasons to be positive and optimistic about what lies ahead. No object or algorithm is ever either good or evil in itself. It’s how they’re used that matters. GPS was invented to launch nuclear missiles and now helps deliver pizzas. Pop music, played on repeat, has been deployed as a torture device. And however beautifully made a garland of flowers might be, if I really wanted to I could strangle you with it. Forming an opinion on an algorithm means understanding the relationship between human and machine. Each one is inextricably connected to the people who build and use it.

  This means that, at its heart, this is a book about humans. It’s about who we are, where we’re going, what’s important to us and how that is changing through technology. It’s about our relationship with the algorithms that are already here, the ones working alongside us, amplifying our abilities, correcting our mistakes, solving our problems and creating new ones along the way.

  It’s about asking if an algorithm is having a net benefit on society. About when you should trust a machine over your own judgement, and when you should resist the temptation to leave machines in control. It’s about breaking open the algorithms and finding their limits; and about looking hard at ourselves and finding our own. About separating the harm from the good and deciding what kind of world we want to live in.

  Because the future doesn’t just happen. We create it.

  Power

  GARRY KASPAROV KNEW exactly how to intimidate his rivals. At 34, he was the greatest chess player the world had ever seen, with a reputation fearsome enough to put any opponent on edge. Even so, there was one unnerving trick in particular that his competitors had come to dread. As they sat, sweating through what was probably the most difficult game of their life, the Russian would casually pick up his watch from where it had been lying beside the chessboard, and return it to his wrist. This was a signal that everybody recognized – it meant that Kasparov was bored with toying with his opponent. The watch was an instruction that it was time for his rival to resign the game. They could refuse, but either way, Kasparov’s victory was soon inevitable.1

  But when IBM’s Deep Blue faced Kasparov in the famous match of May 1997, the machine was immune to such tactics. The outcome of the match is well known, but the story behind how Deep Blue secured its win is less widely appreciated. That symbolic victory, of machine over man, which in many ways marked the start of the algorithmic age, was down to far more than sheer raw computing power. In order to beat Kasparov, Deep Blue had to understand him not simply as a highly efficient processor of brilliant chess moves, but as a human being.

  For a start, the IBM engineers made the brilliant decision to design Deep Blue to appear more uncertain than it was. During their infamous six-game match, the machine would occasionally hold off from declaring its move once a calculation had finished, sometimes for several minutes. From Kasparov’s end of the table, the delays made it look as if the machine was struggling, churning through more and more calculations. It seemed to confirm what Kasparov thought he knew; that he’d successfully dragged the game into a position where the number of possibilities was so mind-bogglingly large that Deep Blue couldn’t make a sensible decision.2 In reality, however, it was sitting idly by, knowing exactly what to play, just letting the clock tick down. It was a mean trick, but it worked. Even in the first game of the match, Kasparov started to become distracted by second-guessing how capable the machine might be.3

  Although Kasparov won the first game, it was in game two that Deep Blue really got into his head. Kasparov tried to lure the computer into a trap, tempting it to come in and capture some pieces, while at the same time setting himself up – several moves ahead – to release his queen and launch an attack.4 Every watching chess expert expected the computer to take the bait, as did Kasparov himself. But somehow, Deep Blue smelt a rat. To Kasparov’s amazement, the computer had realized what the grandmaster was planning and moved to block his queen, killing any chance of a human victory.5

  Kasparov was visibly horrified. His misjudgement about what the computer could do had thrown him. In an interview a few days after the match he described Deep Blue as having ‘suddenly played like a god for one moment’.6 Many years later, reflecting on how he had felt at the time, he would write that he had ‘made the mistake of assuming that moves that were surprising for a computer to make were also objectively strong moves’.7 Either way, the genius of the algorithm had triumphed. Its understanding of the human mind, and human fallibility, was attacking and defeating the all-too-human genius.

  Disheartened, Kasparov resigned the second game rather than fighting for the draw. From there his confidence began to unravel. Games three, four and five ended in draws. By game six, Kasparov was broken. The match ended Deep Blue 3½ to Kasparov’s 2½.

  It was a strange defeat. Kasparov was more than capable of working his way out of those positions on the board, but he had underestimated the ability of the algorithm and then allowed himself to be intimidated by it. ‘I had been so impressed by Deep Blue’s play,’ he wrote in 2017, reflecting on the match. ‘I became so concerned with what it might be capable of that I was oblivious to how my problems were more due to how badly I was playing than how well it was playing.’8

  As we’ll see time and time again in this book, expectations are important. The story of Deep Blue defeating the great grandmaster demonstrates that the power of an algorithm isn’t limited to what is contained within its lines of code. Understanding our own flaws and weaknesses – as well as those of the machine – is the key to remaining in control.

  But if someone like Kasparov failed to grasp this, what hope is there for the rest of us? Within these pages, we’ll see how algorithms have crept into virtually every aspect of modern life – from health and crime to transport and politics. Along the way, we have somehow managed to be simultaneously dismissive of them, intimidated by them and in awe of their capabilities. The end result is that we have no idea quite how much power we’re ceding, or if we’ve let things go too far.

  Back to basics

  Before we get to all that, perhaps it’s worth pausing briefly to question what ‘algorithm’ actually means. It’s a term that, although used frequently, routinely fails to convey much actual information. This is partly because the word itself is quite vague. Officially, it is defined as follows:9

  algorithm (noun): A step-by-step procedure for solving a problem or accomplishing some end especially by a computer.

  That’s it. An algorithm is simply a series of logical instructions that show, from start to finish, how to accomplish a task. By this broad definition, a cake recipe counts as an algorithm. So does a list of directions you might give to a lost stranger. IKEA manuals, YouTube troubleshooting videos, even self-help books – in theory, any self-contained list of instructions for achieving a specific, defined objective could be described as an algorithm.

  But that’s not quite how the term is used. Usually, algorithms refer to something a little more specific. They still boil down to a list of step-by-step instructions, but these algorithms are almost always mathematical objects. They take a sequence of mathematical operations – using equations, arithmetic, algebra, calculus, lo
gic and probability – and translate them into computer code. They are fed with data from the real world, given an objective and set to work crunching through the calculations to achieve their aim. They are what makes computer science an actual science, and in the process have fuelled many of the most miraculous modern achievements made by machines.

  There’s an almost uncountable number of different algorithms. Each has its own goals, its own idiosyncrasies, its clever quirks and drawbacks, and there is no consensus on how best to group them. But broadly speaking, it can be useful to think of the real-world tasks they perform in four main categories:10

  Prioritization: making an ordered list

  Google Search predicts the page you’re looking for by ranking one result over another. Netflix suggests which films you might like to watch next. Your TomTom selects your fastest route. All use a mathematical process to order the vast array of possible choices. Deep Blue was also essentially a prioritization algorithm, reviewing all the possible moves on the chessboard and calculating which would give the best chance of victory.

  Classification: picking a category

  As soon as I hit my late twenties, I was bombarded by adverts for diamond rings on Facebook. And once I eventually got married, adverts for pregnancy tests followed me around the internet. For these mild irritations, I had classification algorithms to thank. These algorithms, loved by advertisers, run behind the scenes and classify you as someone interested in those things on the basis of your characteristics. (They might be right, too, but it’s still annoying when adverts for fertility kits pop up on your laptop in the middle of a meeting.)

  There are algorithms that can automatically classify and remove inappropriate content on YouTube, algorithms that will label your holiday photos for you, and algorithms that can scan your handwriting and classify each mark on the page as a letter of the alphabet.