The Robots Are Coming!

by Andres Oppenheimer

  As of 2018, REX was employing twenty-two human real estate agents in its offices in New York, Austin, Denver, and California to help clients navigate the company’s website or show a property in person, CNBC reported. But these agents are often accompanied by a robot programmed to answer some seventy-five questions that an agent would be hard pressed to know off the top of his or her head. If a client wants to know when was the last time that the property’s roof was replaced or the pool was resurfaced, the robot is there to provide answers. How long will it be before the robots are showing houses on their own, without their human companions?

  AUTOMATED GUARDS AND INSPECTORS

  Even security guards at department stores, banks, and apartment buildings are being automated. Advances in sensors and their abilities to detect anomalies are increasingly replacing human surveillance systems. But what, you might ask, about the security guards who sit surrounded by screens streaming feeds from closed-circuit TV cameras and monitor what’s happening on these screens? We are still seeing such guards in supermarkets and office buildings, but that isn’t likely to last much longer. Computers have a distinct advantage over humans because—unlike people—intelligent machines don’t take breaks, have lapses in concentration, or spend time chatting on their cell phones. Plus their memories are easily searchable.

  And what about the inventory inspectors who, with little notebooks in hand, visit vending machines at the end of the day to determine how much merchandise has been sold and needs to be replaced? This task has ceased to exist in many places. Thanks to the so-called Internet of Things—the new technology that allows things to be connected up with one another online—today’s vending machines have sensors that detect for themselves how many items are missing; they then communicate that information directly to the distributor without the need for human intervention. And in a matter of years, the delivery truck will be an autonomous vehicle, and the person in charge of restocking the vending machine won’t be a flesh-and-blood human but a robotic arm that will come out of the truck.

  The same will be true of water, electricity, and machine inspectors. New automatic sensors connected to the Internet will replace human inspectors of all kinds. In cities like Doha, São Paulo, Beijing, and Barcelona, for example, water pipes have already been equipped with sensors that track the flow and detect leaks, allowing them to save huge amounts of money. In the near future, it will be possible to place cheap sensors on light poles near sidewalks and in other public places to detect sounds and images, which will reduce the need for actual guards, inspectors, and police officers.

  WILL ROBOTS REPLACE DOCTORS?

  Even medical doctors will have to get used to working with robots. According to Silicon Valley multimillionaire innovator Vinod Khosla, technology will replace 80 percent of the work done by doctors, starting with diagnostics. Today, in some of the best U.S. hospitals, many diagnostic tests are being performed by IBM’s Watson supercomputer, which can analyze far more data than any single doctor. While a physician makes his decisions based on his own experience and acquired knowledge, the Watson computer at Memorial Sloan Kettering Cancer Center in New York makes its diagnostics based on the medical histories of 1.5 million patients and 2 million pages of academic articles published in scientific journals.

  What doctor could compete with that? Watson can take the symptoms, genetics, and medical histories of each patient and compare them with hundreds of thousands of similar cases, some successful, others not, and make a decision based on solid statistics as to what treatments work best in each and every case. Who should we trust more, a doctor with the experience of a few thousand patients, or a computer with access to millions of clinical cases? If our grandchildren or future anthropologists come across the television show House—in which the main character, a diagnostic medicine specialist named Dr. Gregory House, imposes his diagnoses on his colleagues based mostly on his instincts and beliefs—they may raise their eyebrows and scream, “What? Is that how medicine was practiced in the early twenty-first century?”

  As we will see in the chapter dealing with the future of medicine, the gurus of medical technology agree that medicine is no longer a practice based on the personal experience of professional doctors being able to sniff out a solution, but a science based on data provided by smart machines. Instead of doctors listening to the heart with a stethoscope, eyes closed so they can concentrate, and measuring blood pressure with a rubber cuff around the arm, as has long been the practice, digital devices with more precise technologies will be used. Some of them are already on the market. And when it comes to surgeries, which in many cases are already being performed by robots, they will be handled almost exclusively with robotic hands, which tremble much less than human ones.

  SIRI, I HAVE A HEADACHE

  New technologies will allow us to perform growing numbers of medical tests at home, resulting in fewer trips to a doctor’s office or a hospital. Many of us are already using search engines like Google, or the personal assistants on our phones, like Siri, or virtual assistants such as Alexa or Cortana, to ask for medical advice. And thanks to artificial intelligence, these personal assistants will soon be able to respond to medical queries as well as or even better than many doctors. Alexa, Amazon’s virtual assistant, a little intelligent loudspeaker that I have in my own home, can already give me precise instructions on what to do when I have anything from a headache to a heart attack, or how to perform cardiopulmonary resuscitation on a person in an emergency.

  As we move forward, we will be increasingly relying on Siri, Alexa, or their successors for answers to our medical questions. If our children have a fever, the first step won’t be to go to the doctor’s office. Instead, it will be to consult a robot, which—based on the information we give it, plus what it receives from its sensors—will decide whether to recommend an aspirin or a trip to the nearest emergency room. What tasks will be left to the doctors? I’ll talk more about that later in the book, but here’s a preview: it will have a lot to do with explaining intelligent machines’ diagnoses to patients and holding people’s hand during the process.

  As robots become more highly skilled and less expensive, even anesthesiologists will be replaced by robots. In the United States, an automatic sedation system called Sedasys, from Johnson & Johnson, was approved in 2013 for patients undergoing a colonoscopy. The system allowed one doctor to monitor up to ten patients simultaneously, instead of a one-to-one doctor-to-patient ratio, thus dramatically reducing hospital costs. While sales of Sedasys were recently discontinued, among other things because of pricing issues and anesthesiologists’ objections, it’s only a matter of time until a new version of it becomes a must for most hospitals.

  What about pharmacists? you may be asking yourself. They will have to reinvent themselves as much or more so than doctors. For the past several years, the University of California at San Francisco has been using a pharmaceutical robot that has filled over 2 million prescriptions without a single mistake, and its example is being followed by many hospitals across the United States. By comparison, average human pharmacists make a mistake in about 1 percent of prescriptions they fill.

  ROBOTS WILL BE LAWYERS AND EVEN JUDGES

  Robots are handling growing numbers of chores at law firms and are increasingly offering legal advice to the outside world. In 2016, the giant U.S. law firm DLA Piper, with over four thousand attorneys across thirty countries, used a computer program designed by the artificial intelligence company Kira Systems to analyze corporate contracts and propose changes, which were tasks ordinarily assigned to young lawyers who had just joined the firm. Meanwhile, many other firms are starting to use online legal services platforms such as LegalZoom and Rocket Lawyer to collect data, something that used to be done by young associates or paralegals.

  These same online platforms are also offering customer services such as writing deeds, commercial contracts, and even divorce papers. In other words, th
e algorithms they use are skipping right past the lawyers, resulting in automated legal services that end up being much cheaper for their clients. Just as online websites have replaced many travel agencies, legal websites are replacing many attorneys who used to handle routine tasks like preparing real estate rental contracts or prenuptial agreements.

  THE CASE OF THE ISRAELI APPEALS COURT JUDGES

  Even judges—members of a highly prestigious profession demanding skills we wouldn’t normally associate with computers, like the ability to use good judgment and common sense—are at risk of being replaced by much more efficient algorithms. According to their advocates, these computer programs are not prejudiced and can issue far more impartial verdicts than humans. This is no joke: a study done by Professor Shai Danziger, now at Tel Aviv University in Israel, which examined the rulings by eight Israeli judges over a ten-month span, found that they handed down more generous judgments after having lunch.

  According to this study, published in the American journal Proceedings of the National Academy of Sciences in 2011, the eight judges—who were in charge of accepting or rejecting prisoners’ requests for paroles or sentence reductions—approved roughly two thirds of the requests they handled early in the morning. But later in the morning, the number of applications they approved fell dramatically…until lunchtime. After eating, the judges again tended more toward clemency.

  Danziger and his coauthors concluded that the judges’ grumpiness increased as time passed since breakfast, affecting their verdicts. They speculated that it could be attributed to two factors. The first had to do with the judges’ blood sugar levels, which decreased progressively as the hours went on since their last meal. The second key factor didn’t have to do with the amount of time that had passed, but rather the number of cases over which they had presided. According to this latter explanation, decision-making is an arduous and exhausting task, and it was the caseload that made them crankier and led them to issue harsher verdicts. Whichever the case may be, the study concluded that the judges hadn’t shown a racial or gender-based bias in rendering their rulings. The primary constant among their verdicts was simply the time of day. In other words, robotic judges could be much more impartial than humans.

  ROBOTIC PROFESSORS ARE ALREADY TEACHING CLASSES

  When it comes to education, there are already robots in the marketplace that promise to carry out a number of tasks currently performed by teachers. According to their manufacturers, these robots have several advantages over human educators, including the fact that they have infinite patience when it comes to explaining things. Unlike human teachers, who tend to get exasperated after several unsuccessful attempts to explain something to their students, a robotic teacher can explain things in hundreds of different ways. If a student doesn’t follow a particular explanation, the robot can move on to the next one, and the next one, and so on.

  One of the most charming robot professors I met, and who I’ll describe in detail later in the book, was Professor Einstein, a smart doll standing fifteen inches tall with the face and white hair of the world-known Nobel Prize–winning physicist. The doll has multiple facial expressions and gestures, and when it says something mischievous, it even sticks out its tongue. The creators of Professor Einstein claim the robot is able to explain physics and mathematics to students better than most human teachers because it has dozens of alternative teaching methods, depending on whether the student learns better verbally or visually, from playing or by solving problems. In other words, if the student doesn’t understand the lesson as it’s being explained orally, Professor Einstein can explain it in a visual manner by showing a video or illustration. And while robotic teachers aren’t about to replace their human counterparts, they can at least help children with their homework. If young people now prefer interacting with tablets rather than human beings in a restaurant, why should it be any different when it comes to a classroom or doing their homework?

  BIONIC SOLDIERS: HALF HUMANS, HALF ROBOTS?

  The field of robotics is advancing at such a pace that even cyborgs—those half-human, half-robot warriors we saw in Arnold Schwarzenegger movies—could come true. I saw evidence of this with my own eyes when I interviewed Dr. Hugh Herr, also known as the Bionic Man. Herr is currently serving as director of the Biomechatronics group at MIT’s Media Lab, and I had him as a guest on my television show. While I knew his personal story well (he had been the cover story in many major magazines) I had never thought about the potential impact his discoveries might have for the future of humanity.

  Herr had been an avid mountain climber from a very young age, and was considered something of a child prodigy in the climbing world. When he was seventeen, he and a partner were climbing in the northeastern United States when they became trapped for four days during a blizzard; he suffered severe frostbite, and doctors were forced to amputate both his legs below the knee. When he was fitted with artificial limbs, Herr couldn’t understand how, with all the new technology available to us, the prosthetics being used were still nothing more than the same stiff pieces of metal that had been used for centuries. So he set out to create his own prosthetic limbs: he studied mechanical engineering at MIT and completed his doctorate in biophysics at Harvard. After that, he set up his laboratory on the MIT campus and began exploring new technologies for increasing the physical capabilities of human beings, starting with his own body. His dream was, quite literally, to be able to climb mountains again.

  Today Herr uses his own bionic legs to climb everything from stairs to mountains more easily and more powerfully than anyone with natural limbs. While spring-based prostheses like those used by South African sprint runner Oscar Pistorius are passive extensions of the human body, Herr’s legs are intelligent, with their own power supply, six computers, and twenty-four separate sensors. They allow him to walk and run much more naturally and powerfully. During our interview, he told me that while the rest of us may keep a dozen or so pairs of shoes in our closet, he has more than a dozen different pairs of legs that he uses for walking, running, and every sort of rock or ice climbing imaginable.

  According to Herr, his inventions will allow us to put an end to physical impossibilities and delete the word disabled from our vocabulary. With bionic arms or legs, people with physical limitations can be as able as any other, and perhaps even more so. “You cannot, with a straight face, label me as disabled,” he told me. “My ultimate dream for the world is that disability is no more, and I believe that this will come to pass in this twenty-first century…I think fifty years from now, most disabilities will have been eliminated. Blindness, paralysis, severe depression, all these disabilities of today will be eliminated.”

  But isn’t all this potentially dangerous? I asked. Couldn’t it lead to a world of half-human, half-robot supersoldiers who could overpower the world with their bionic arms and legs? Herr shrugged, and replied that since the Stone Age, the armies with more advanced technology have held an advantage over those with more primitive tools. “I call that ‘augmentation,’ ” Herr explained. “Augmentation is a technological enhancement that enables us to overcome our normal physiological levels. You know, augmentation technology has been here for a very long time. It’s actually pervasive in society. It’s so pervasive that it’s commonplace, and we often don’t recognize it as such. The bicycle is an example: I can hop on a bicycle and go from point A to point B much faster and with much less energy than by using my own legs. The automobile is another profound example. So there’s augmentation all around us. It’s pervasive. So as we march into the future, there will just be more and more levels and forms of augmentation.”

  THE SOLDIER ROBOTS OF SOUTH KOREA

  In South Korea, I got a bit of a preview of the world Herr was describing to me. While for years the United States has been using unmanned aerial vehicles known as drones to fight terrorists in the Middle East, and is increasingly automating its armed forces, few countries are developing robotic
soldiers as quickly as South Korea. Faced with constant threats from its neighbor to the north, South Korea is thinking about replacing a significant number of troops along the demilitarized zone with robots.

  “Our soldiers are currently hunkered down in a bunker, their rifles aimed permanently north, but they’re freezing to death,” Junku Yuh, director of robotics at the prestigious Korea Institute of Science and Technology (KIST), told me. “But very soon we will be replacing them with robots, which our soldiers will be operating remotely from a heated building. So we will need far fewer soldiers, and if the enemy attacks us, they will be attacking robots, not men.”

  A South Korean arms company has already produced a robotic machine gun turret known as the Super aEgis II, which has a firing range of 2.5 miles and is paired with a loudspeaker whose voice has a range of nearly 2 miles. The sound is delivered with extraordinary precision, issuing a warning to a potential target advancing from the north before they are fired upon. “Turn back,” it says in rapid-fire Korean. “Turn back or we will shoot.” And the we is important, because there is a human being supervising the robotic turret from a remote location who must manually enter a password and give the order to shoot before the machine gun will open fire. However, as the manufacturer says, human supervision is simply there to calm the public, because the robot is perfectly capable of fulfilling its mission without the need for human approval. The most important thing about the robot—which has already been sold to the armed forces of Dubai, Abu Dhabi, and Qatar—is that it identifies potential threats through cameras and sensors that the company claims are so acute that they can even detect whether a suspect is carrying explosives underneath his clothing. That’s something that robots can do much better than humans, the company says.