George and the Ship of Time

by Lucy Hawking

  In the past, adding more and more people to our cities has resulted in extensive pollution, traffic jams, housing shortages, and huge demands on services. City planners of the future will need to consider how to manage these issues if they want to make cities great places, rather than places that we only tolerate because that’s where our jobs are.

  Where will we live, work, and go to school in these cities of the future? What will those experiences be like? Will we have robot butlers? Will we have to work, at all, or will everything be done by robots?

  As we have seen since the beginning of the Industrial Revolution, many jobs that had previously been undertaken by people have been mechanized. There is no reason to think that this trend will change in the future. But there will continue to be a need for people to design the machines and the robots that we will use to do many of these tasks. And lots of things can’t be done by machines: creative jobs such as writing books and creating art, designing buildings or computer games. These areas will continue to need people and their ideas. Maybe we’ll work fewer days per week in the future, but people could then spend more time with their families, helping their communities or having fun.

  No matter what jobs we may be doing, we’ll still need a place to do this work. Although technology continues to develop so that a lot of our work can be done from anywhere with an Internet connection, many people still choose to go to offices or other spaces where they can collaborate with others. So we’re likely to continue to want some kind of building in which to talk to each other and share ideas. As more and more high-rise office buildings are being developed around the world, it is unlikely that our skylines will change completely in the future, but these offices are likely to be designed to be attractive places to work. There is increasing demand for outdoor spaces in office towers, so although the skyline may not change, it is likely to look a lot more green than it currently does, with terraces, roof gardens, and green walls.

  Different cities have already shown different approaches to where we live—some cities have lots of houses, while others have lots of apartment buildings. As cities become more densely populated, it is likely that housing will need to be intensified—meaning that more people will need to live in the same small area. City planners will need to consider how to develop additional housing—and make it affordable for all types of people to live there—to meet the needs of growing populations.

  Whatever the outsides of our houses look like, however, technological changes are likely to make the insides different from today. Many of the devices that currently exist will continue to develop to make our lives easier: smart devices should be able to tell us how much energy we use so that we can use less; other technology can turn on our music or let the cat out. And 2017’s Alexa will likely develop into a full-scale robot butler to take care of many more of our household chores, just like in The Jetsons.

  Schools will also take advantage of changes in technology. Will we need to go into a school building? For the same reason that people prefer to go to offices, children of the future will probably still attend school and teachers will still be humans rather than robots. But technology will develop in ways that allow for virtual and augmented reality that allows children to “go” to the rain forest or experience the French Revolution or the Roman Empire more than we can today.

  So if we know what we’re doing in the future with regard to work and school and home, what else needs to be considered so that our cities can be amazing places to live? The big issues affecting cities today are likely to continue to be the big issues of the future: transportation and our environment.

  If our cities are getting bigger and more populated, it will be harder for people to move around easily in cars. Public transportation will be key to minimizing the number of people stuck in traffic. Planners will need to consider if more underground trains make sense, or if alternative transport solutions are preferable. Driverless vehicles are likely to be more and more prominent, but will these cars create more traffic or less? We will need to come up with solutions to manage driverless vehicles more effectively, rather than this simply resulting in more cars on the road.

  Will we need to care about traffic and public transport at all if there are flying cars? Probably even more so. Just because cars can fly, it doesn’t mean that traffic and pollution will go away. Combine flying cars with delivery drones and airplanes and helicopters, and there could be some very busy and polluted skies!

  Transport uses a lot of energy, which has an impact on the environment. Putting millions of people in a single city location is going to have an impact on the environment as they cook, turn on lights, heat and/or cool their homes, charge their phones, use computers and TVs, and travel around. All these things require energy, and energy consumption has historically had a negative effect on the environment.

  Many city governments are now looking at how they can lessen their impact on the environment, especially reducing pollution that could be harming their residents. Effort will need to go into reducing energy consumption and finding environmentally friendly energy solutions to deliver our needs. More and more electricity is generated through renewable and low-carbon means, but really innovative solutions may be the best ways to create the energy we need for the future: hydrogen cars could replace existing gas and diesel cars, and their only exhaust would be water vapor rather than carbon dioxide. Technology could be developed that turns human power, generated by walking or cycling, into electricity. Or that turns our homes, offices, and schools into energy generators in some way, allowing us each to self-generate our own requirements. Perhaps you, in the future, will be one of those who will design technology like this, or will help in planning and building our cities of the future.

  A strong vision of what we want these cities to be like will be required so that we can capture all the benefits that technology could give us in our lives. Do you have this vision? I began by imagining my city of the future based on a TV cartoon series. What sort of city can you imagine?

  Maybe not flying cars, but hopefully lots of robot butlers!

  Artificial Intelligence

  by Dr. Demis Hassabis

  What does it mean to be intelligent? Most often in daily life, the term is used to describe how well someone does at math, writing, or another academic subject, but there is a more basic definition. At its core, intelligence means the ability to achieve goals in a wide variety of environments. Sometimes your goal might be solving a math problem, but other times it might be something much simpler that we usually take for granted: describing the weather, playing a computer game, or using a knife and fork to eat a meal. Although we don’t usually think of these as particularly challenging tasks, they actually involve a tremendous amount of computer power, and it is remarkable that our brains are able to do so many different types of activity so well.

  Intelligence is what makes humans exceptional when compared to other animals: by looking at the world around us and thinking about how it works, we have built tools, societies, and civilizations to help us achieve our goals. In the span of a few tens of thousands of years—the blink of an eye relative to the history of life on Earth—humans have used our intelligence to make incredible progress: discovering electricity, building skyscrapers, curing diseases, mastering flight, and even sending people to the Moon and launching probes past the limits of our solar system. Our intellect that has powered these achievements is unlike anything else that has ever taken place on this planet, and possibly unlike anything else in the entire Universe.

  Imagine if we had intelligent machines that could help us discover even more new inventions and answer even more questions! This is exactly the goal of artificial intelligence, or AI.

  For a long time computers have been excellent at some tasks, such as math and logic, but have not been nearly as flexible as human minds. Activities that we find easy—like identifying different animals or carrying on a conversation—have generally been incredibly difficult to automate. But as computers ha
ve become faster, people have discovered new ways of programming them that have unlocked some of these abilities. Today, a number of the world’s most brilliant scientists are working on designing new programs (or algorithms) that will enable computers, like humans, to apply intelligence to accomplish goals in a wide variety of environments. This is AI.

  The most exciting area of AI research at the moment is called machine learning. Machine learning takes a different approach from normal computer programming: instead of giving the computer precise, step-by-step instructions, machine-learning researchers write learning algorithms that allow computers to observe the world around them and figure out answers for themselves. For instance, instead of writing a program that tells a computer that a cat has two eyes, four paws, and whiskers, a machine-learning researcher might write a learning algorithm and then simply show it a lot of different pictures of cats. Over time, the algorithm will learn from these examples to identify cats for itself. This is very similar to how we teach human children: we might simply say, “This is a cat,” or “This is a dog,” and let the child figure out independently what the differences are between cats and dogs.

  One of the most wonderful and powerful aspects of machine learning is that it is much more adaptable than regular programming. For instance, we could take the same algorithm that we used to identify cats and train the computer to identify all sorts of different animals. We could also use it to recognize faces, cars, buildings, trees, and pretty much anything else. This saves us a huge amount of effort because we don’t need to write specific programs for each problem! Because the algorithms are general-purpose, they can be used in all sorts of different situations.

  Another benefit of learning algorithms is that, unlike normal computer programs, they can discover new facts and strategies that we did not know when we created them. For example, just recently an AI program called AlphaGo defeated the best player in the world at an ancient Chinese board game called Go. Go is sort of like chess, but much, much more complicated: it has more possible board positions than the number of atoms in the entire Universe! This makes the game very difficult, and the world’s best players spend their entire lives honing their skills and trying out new tactics. AlphaGo is a machine-learning program that, much like human players, learned to play the game by experimenting over time with lots of different moves and seeing which ones worked best. This meant that it discovered some novel strategies that no human player had ever used, so it not only won the game but also taught human Go players worldwide about powerful new techniques—this could never have happened with an algorithm that had been programmed conventionally with step-by-step directions. AlphaGo was a major milestone for AI because it demonstrated the power of learning algorithms to make their own discoveries in very complex domains.

  Of course, we have not yet built anything nearly as flexible or capable as the human mind; there are lots of tasks that we humans find easy but even the best AI algorithm remains unable to do. But over the past few years machine learning has made tremendous progress. In addition to playing Go and identifying people and animals, machine-learning programs have translated languages, improved energy efficiency, and made medical advances, to list just a few of the many astounding recent examples of AI.

  All this, however, is just the tip of the iceberg. Ultimately, AI scientists hope to achieve artificial general intelligence (AGI)—an AI algorithm capable of doing anything the human brain can do—which would be invaluable in helping scientists conduct important research and uncover new truths. Having AGI will usher in a new age of tremendous scientific discovery: just as humans have made amazing progress over the past few thousand years by applying our own intelligence to various problems, imagine what we can accomplish if we can combine that intelligence with the power of AI! We might be able to cure most diseases, solve difficult problems like climate change, and discover miraculous new materials that could improve space travel.

  This is a very exciting time for machine learning. It seems as though almost every day brings a new discovery that inches us closer to artificial general intelligence. Achieving AGI will be a huge breakthrough for mankind—something on the same level as the Moon landing or the creation of the Internet. Over the course of human history we have built many tools—ranging from hammers and shovels to telescopes and microscopes—but none of them have had the same potential as AI for revolutionizing almost every aspect of human life.

  Of course, nobody can say for sure how far we are from AGI. But at the speed that the field is progressing, it could happen within our lifetimes, in which case right now we are standing on the brink of a world-changing discovery, gazing into a future bursting with possibilities. There has never been a more thrilling time to be alive!

  It is a fascinating and hugely exciting area to work in. In the years to come, maybe you—as one of the current generation of young people for whom computers are a familiar part of everyday life—will be one of the programmers who develop AGI further and further and use your skills to help our society achieve truly amazing things!

  Robot Ethics

  by Dr. Kate Darling

  Is it okay to be mean to a robot?

  We all know that robots are just machines that are programmed to do things. You can’t hurt their feelings and they don’t experience pain like humans and animals. But . . . if verbal or physical violence toward robots still feels wrong to you, that’s not crazy!

  There’s an interesting phenomenon in human psychology called anthropomorphism. It means that we project human qualities and emotions onto nonhumans. If you’ve ever thought that a stuffed animal looked sad because it was thrown under the bed, or that a dog was smiling happily at you, you’ve experienced anthropomorphism. Dogs certainly have emotions, but they’re harder to read than most people think! We sometimes take cues from animals and objects and imagine that they feel the same as a human would. And even though we may be wrong about what we’re imagining, it’s a pretty natural thing to do—evolutionarily, it’s how we try to make sense of, and relate to, other beings and things.

  It turns out that we anthropomorphize robots a lot. Robots combine two factors that evolution has taught us to respond to: physicality and movement. We’re very physical creatures, and our brains are hardwired to see life in certain types of movements. So, if we see a robot in our physical space that seems to be moving around all by itself, part of our brain thinks that the robot is doing things intentionally. And that makes it easy to imagine that the robot has goals and emotions. That’s why a lot of us feel sorry for a robot when it gets stuck somewhere, even though the robot really doesn’t care at all if it’s stuck!

  Some robots are specially designed to target this instinct. Have you seen Star Wars? Just like R2D2 and other robots in Star Wars, we can make real robots that use sounds and movements and other cues that we automatically associate with living things. A lot of children and adults enjoy playing with these robots because it’s so easy to imagine that they’re alive. And this imagination can even be used to help people in health and education. For example, robot animals can be pets for lonely or sick people who are allergic to real animals. Teachers can use robots as friendly and engaging sidekicks, to make learning more fun. Some robots are already really good at reminding people to take medicine, or comforting them, or motivating them to learn new languages. And these robots are helpful because people treat them like living things instead of like devices. It’s more fun to talk to a robot than to a toaster or a computer!

  Maybe someday soon you’ll have a robot helper at home. But before you tell your robot all your secrets, here’s something to keep in mind: it’s important to know a little bit about how the robot works, what purpose it serves, and what data it collects about you. For example, is the robot recording what you say? If you tell it something personal, can somebody else get that information? Most companies that sell robots probably just want you to have a cool robot, but some of them may want to collect your data to sell to other big companies. Or they may
have some other idea to make more money using the robot. After all, robots are machines made by people, so they do what their creators want them to do. That’s not always a bad thing. It’s just a good idea to take a moment to ask: who made this robot, and why?

  In the future, robots will be in a lot of places and made for many different tasks. Some robots will be programmed to act as if they have feelings. And that brings us back to the question: is it okay to be mean to a robot? If robots don’t really have feelings, it’s not as bad as being mean to animals or people. But if you’re nice to robots, you’re not being silly. In fact, it may mean that you have a lot of empathy. Scientists like me have been researching the ways in which we treat robots like they’re alive. One of our questions is whether we can learn anything about a person from how they act toward a robot. So far, we think that people who feel empathy for robots have a lot of empathy for other people too. So before being mean to a robot, consider this: if you’re a kind and caring person, that may not matter to the robot, but it sure matters to you and others!

  The Internet: Privacy, Identity, and Information

  by Dave King

  Have you ever thought about who can see what you do on the Internet or how long the messages you write will last?

  The Internet is made up of many, many different computers all interconnected across the world. We tend to access the Internet through our mobile phones and other devices but some computers are designed specifically to store the information we all put on the Internet. These computers, called servers, host the websites we access. Some of them are in homes and offices, but most are in purpose-built centers run by Internet Service Providers (for short, ISPs). Big companies like Google, Facebook, and Amazon have their own data centers—and networks of machines that each hold huge amounts of data.