Tapia was much smarter than her dinosaur colleagues on the ground floor. Plus she had a sense of humor and was programmed to have conversations unrelated to her functions. When I told her jokingly that she was a very cute robot, Tapia replied, “Thank you very much, I hope you don’t say the same thing to other robots besides me.” Tapia was more comfortable communicating in English than the reptiles at the reception desk. She probably had more experience speaking languages, because robots learn from each and every interaction, adding new data to their memory. She had obviously been at her job longer than the robots downstairs.
When it was time for dinner, I called the desk to ask where I could find the restaurant, only to be informed by the human assistant to the robotic concierges that unfortunately there was none. The hotel had only vending machines that offered warm meals, which were located on the ground floor. Since it was late and very cold outside, I had no choice but to buy a bowl of Japanese noodles from the vending machine, which was not unlike those that sell soft drinks. I put the money into a slot, pressed a button showing an image of the noodles, and a box with my dinner dropped into a small microwave in the lower part of the machine, which automatically turned on, heated the food, and dispensed it for me. The noodles, by the way, were dreadful.
Since I was already in the lobby, I decided to knock on the door of the human assistant to the dinosaurs and to try to interview her. After a question or two about the hotel, I asked her almost in passing how many actual humans worked there. The young woman, Saki Kato, seemed delighted to practice her English, which was pretty good. She told me that there were only two people working that night: the manager and her. Together they were in charge of the entire hotel—a hundred rooms in all—almost all of which were occupied. Maids are there during the day, but almost all other functions, down to cleaning the windows and washing the floors, are handled by robots, Saki told me. “And, well, we also have people to service the robots, because they often break down and have to be reprogrammed,” she added with a mischievous grin.
The next day, once my curiosity was satisfied, I relocated to another hotel in Tokyo. At $400 per night, the Henn na Hotel was quite expensive, and the room was tiny, even by Japanese standards. There was barely enough room to walk between the bed and the wall. And I didn’t like the idea of having another dinner of reheated noodles from a vending machine in the lobby. I concluded that the dinosaur concierge was doing about the same thing as an airport check-in kiosk, though it did make the experience a bit more fun. And Tapia, the egg-shaped robot, was simply a personal assistant with voice recognition software, not unlike Siri on my smartphone.
But while there weren’t any extraordinary technological advances at the hotel, there was little doubt in my mind that the automation of the hotel industry is just around the corner. When I checked out, the dinosaur concierge asked me to return my plastic room key to a slot, said goodbye, and thanked me in Japanese with a bow that made me smile. As I was leaving, I concluded that perhaps the only reason why more hotels haven’t started using robots yet is due to fears that customers would find the machines too cold and lacking in human warmth. But if the Henn na Hotel was able to overcome this challenge by turning the robotic registration into a fun experience through dinosaurs wearing concierge hats instead of metallic robots, others in the industry are likely to follow its steps soon. The Henn na Corporation has announced that it will be opening six new automated hotels in Tokyo and three in Osaka in 2018 and that in the long term, it is planning on opening a hundred more worldwide. It will be difficult for other hotel chains to compete with that or to resist the temptation of running a hundred-room hotel with just two regular employees, a cleaning staff, and a team of robots that can work twenty-four hours a day, take no vacations, and never ask for raises.
ROBOTS WILL BE EVERYWHERE
Soon enough, robots won’t be seen just in hotels but also in the streets, schools, hospitals, law firms, and almost everywhere else. Industrial robots have been used since the 1960s to perform repetitive and relatively simple tasks, especially in the auto industry. But until now they had not spread much beyond the manufacturing floor. Despite the futuristic predictions of cartoons like The Jetsons and science fiction films that have been predicting the emergence of robots as domestic servants, drivers, and even as pets, the science of robotics had remained stagnant for decades.
But now robotics is growing by leaps and bounds, thanks to the fact that robots are getting cheaper and smarter. Artificial intelligence and cloud computing (the massive online database commonly known as the cloud) is allowing each individual robot to learn from the experiences of all others. Before, a robot was an individual machine that carried its own information internally, sharing it with a small group of similar automatons, if at all. But now every robot connected to the cloud has immediate access to an almost unlimited amount of data and the experiences of the entire globe’s robotic population, allowing them to be constantly learning from one another. And that is revolutionizing the world of work.
Ever since the supercomputer Deep Blue defeated world chess champion Garry Kasparov in 1997, robots have been taking down one challenge after another. In 2002, a software program beat the world’s best Scrabble players. In 2010, another program took down some of the greats at the game of bridge. In 2011, IBM’s Watson supercomputer beat two champions of the popular TV game show Jeopardy! And in 2016, AlphaGo, a computer Go program developed by Google’s DeepMind program, made headlines around the world by defeating South Korea’s Go champion Lee Sedol. Until then, the game of Go was considered unwinnable for a machine, because in addition to intelligence, it requires a good amount of intuition and creativity.
Some scientists, including Vernor Vinge, who is also an accomplished science fiction author, have predicted that intelligent machines will surpass all human capabilities by 2023. Others, such as Google’s futurologist and engineering director Ray Kurzweil, predicted that singularity—that moment in time when artificial intelligence surpasses human intelligence—will occur in 2045. With technology advancing at the rate that it is, it would not be surprising for it to happen sometime in between. Recently, a robot named Michihito Matsuda ran for mayor of the Japanese city of Tama, on the outskirts of Tokyo, with the campaign slogan “Artificial intelligence will change Tama City.” Matsuda, who according to press reports was created by two big-tech executives, got 4,000 votes and finished in third place. But how long will it take before another robot runs a better campaign and convinces voters that it can make more levelheaded decisions than a human?
It’s no longer a question of whether technology will surpass human intelligence, but of when. That’s why, when asked how he would prepare for the next chess match against a supercomputer like IBM’s Deep Blue, the great Dutch chess master Jan Hein Donner replied, “I would bring a hammer.”
ROBOTIC RECEPTIONISTS ARE ALWAYS IN A GOOD MOOD
During my recent trip to Japan, I encountered robots in restaurants, retail stores, banks, and office buildings. I met a robot named Pepper at one of the Hamazushi sushi restaurants in Tokyo, and another one at the entrance of a branch of Mizuho, one of the largest banks in Japan. They were welcoming customers and directing them to the appropriate offices, always with a smile and a bow. Robots are also currently being trained to work as car dealers at Nissan branches in Tokyo and to serve customers in shopping centers in a number of cities throughout the country.
Pepper, a doll-like little humanoid robot with big black eyes and a white plastic body, costs less than the annual salary of a Japanese receptionist and provides many more years of useful service. But most important, Pepper can work for twenty-four hours a day, seven days a week, and never asks for vacation time or pay increases. When he operates well—which isn’t always the case, though eventually it will be—he is the ideal employee for companies that want to lower labor costs and avoid union problems. While looking for a place to exchange some currency at T
okyo’s Haneda Airport, I wandered past the entrance of a Bank of Tokyo-Mitsubishi branch, where I was greeted by another robot, named Nao. Like Pepper, Nao was armed with cameras, sensors, and microphones with which he could interact with customers. Nao proudly assured me that he could understand and speak nineteen languages and give customers the daily rate of the yen against any other currency. As the branch manager would later explain to me, Nao was being trained to serve the tourists who would be coming from all over the world for the 2020 Tokyo Olympics.
I spoke into the robot’s microphone, and indeed it was able to recognize the language being spoken to him and reply in kind. In Singapore, a humanoid robot named Nadine is already working as a receptionist and assistant at the Nanyang Technological University, and—using the same technology as other personal assistants on the market, like Siri—she can answer any question someone might ask, whether it’s related to the university or a recommendation for the best restaurant in the area. And as I experienced several times, many of these robots are programmed to have a sense of humor. At the entrance to a men’s clothing store near the Akihabara train station in Tokyo, I saw a robot that sang, danced, and waved its arms, inviting passersby to enter the establishment. Unlike some other human receptionists, he seemed genuinely happy, smiling from ear to ear.
THE END OF RESTAURANT WAITERS?
Automation in the restaurant industry is advancing at full speed. I came across this phenomenon for the first time in Miami in 2016, when I had lunch at one of Panera’s 1,800 U.S. locations. McDonald’s, Wendy’s, Pizza Hut, and virtually every other fast-food chain in America and Europe are using touch screens to allow customers to place their orders electronically instead of having to interact with a human waiter. As soon as I entered the Panera, I saw five small metal towers topped with tablets displaying the menu, images of all the items, and their respective prices. I swiped through the various dishes with my finger, and—as with any online shopping site—I hit okay to confirm my purchase and inserted my credit card.
“Thank you, Andrés,” the machine instantly replied. I was stunned for a moment, wondering how Panera knew who I was, because the tablet had never asked me for my name. But of course it had picked it up from my credit card and treated me as if we were old acquaintances. Then I walked a few steps over to an area where other customers were waiting to pick up their orders, and I realized that there weren’t any employees to let me know when my food was ready. Instead there was a TV screen on the wall with a chronological list of pending orders along with the names of the people who had placed these orders. When a name reached the top of the list, his or her order was ready to pick up. But what amazed me most about the whole process was the fact that there was a restaurant employee standing at the counter, ready to take orders, but very few customers ever approached her. They seemed to prefer to interact with the tablets.
According to industry spokespeople, the main reason restaurants are becoming more and more automated isn’t to save costs but to satisfy their customers: growing numbers of people—especially the younger generations—would rather place their orders through tablets or on their cell phones. I asked several industry representatives why they preferred this to interacting with a person, and the reason was simple: young people don’t to want to spend time waving a hand or trying—often unsuccessfully—to make eye contact with a waiter to get his or her attention. Why go through all that trouble when you can simply order your food electronically? Why sit at a table for ten minutes for a waiter to bring you your bill when you can swipe a card and pay it directly whenever you’re ready? All of this is causing many restaurants to reduce their number of waiters, whose job is now mostly limited to delivering food to customers’ tables. And soon enough, even that could change, with robots or conveyor belts performing that task.
While in Japan, I ate at a number of sushi restaurants that have neither hostesses, waiters, nor cooks. Even the chef who prepares the sushi is a robot. And judging by the success they’re having, they will likely be joined by restaurants serving other kinds of food. Japanese customers, attracted by the low prices and the speed at which automated restaurants operate, are turning to them in ever-growing numbers, often without even realizing that the chef preparing their sushi is a machine. Hamazushi, one of several automated sushi restaurant chains, already operates in 454 locations across Japan. Customers line up in front of a robot hostess who assigns them a table, and then they sit next to a conveyor belt that brings the sushi plates to each and every table. Each of these tables is topped with a tablet computer displaying the menu, a hot water tap for making tea, place settings, chopsticks, and several varieties of soy sauce.
When customers sit down, they choose their preferred dish—there is a picture of all the options, along with their respective prices—and after a few minutes, the tablet bleeps out a little tune and flashes a message that reads “Your order is about to arrive.” And indeed, within a few seconds, a plate with a sign indicating the specific table arrives on the conveyor belt. When their meal is over, customers swipe their credit card through the tablet, and leave.
JENKINS: HALF OF ALL BANK EMPLOYEES WILL DISAPPEAR
Just as robots are threatening the jobs of hotel and restaurant industry workers, the Internet and artificial intelligence are challenging bank employees by creating new ways of investing, lending, and transferring money. Virtual banks, such as Betterment.com, have no physical headquarters, which allows them to reduce costs and charge lower fees than traditional banks. And while these virtual banks are handling more and more private investments, they also face growing competition from other online payment platforms that are increasingly performing routine transactions such as money transfers that, until recently, were carried out only by brick-and-mortar banks. Growing numbers of people are moving money around through their computers and cell phones, and many of the companies that are performing these transactions, including Square, Google Wallet, Apple Pay, and Venmo, are not—nor were they ever—in the banking industry.
Former Barclays CEO Antony Jenkins shook the banking industry to the core in 2015 when he predicted that by 2025, banks will have cut their number of branches—and the number of employees—in half. Jenkins, who had already begun executing a layoff plan for some 19,000 employees when he left the bank, said in a speech that the “Uber moment” had come for the financial world. He said technology is “an unstoppable force,” and predicted that many of the large, traditional banks would either merge or disappear in the coming years because they wouldn’t be able to compete with the new virtual banks or payment systems like PayPal in the United States or Alipay in China.
He wasn’t exaggerating. In northern Europe, the number of bank branches has already started to plummet. Whereas in 2004 there were twenty-five banks per 100,000 citizens in Scandinavian countries, that number dropped to seventeen by 2014, and is projected to fall to eight by 2025, according to predictions by the World Bank and a Citi GPS report. In the United States, the number of bank branches is expected to fall by 33 percent during that same period, and by 45 percent in Latin America, according to the Citi GPS report.
“The return on having a physical network is diminishing,” says Jonathan Larsen, Citigroup’s former global head of retail and mortgage, in that report.
Branches and associated staff costs make up about 65% of the total retail cost base of a larger bank and a lot of these costs can be removed via automation. The pace of staff reductions so far has been gradual (~2% per year or ~11–13% from peak levels pre-crisis). We believe there could be another 30% reduction in staff between 2015 and 2025, shifting from the recent 2% per year decline to 3% per year, mainly from retail banking automation.
Bank CEOs are asking themselves: Why maintain a network of brick-and-mortar banks with human tellers to make cash deposits and handle checks when these transactions can be easily automated and growing numbers of clients are choosing to go paperless? Cash—the bank�
��s raw material, which was already affected by the advent of credit cards—is becoming even more obsolete in the age of electronic payments. Some countries, like Denmark, are seriously considering eliminating cash altogether. In Europe’s Nordic countries, even beggars have gone online: since they know most people don’t carry cash anymore, they ask for transfers to their cell phones. “Could you transfer me a few cents?” they ask.
REAL ESTATE AGENTS REPLACED BY ALGORITHMS
In California, more and more home buyers are looking for properties using algorithms, which charge a 2 percent commission on each sale, rather than real estate agents, who charge 6 percent. REX Real Estate Exchange, an automated real estate company, uses artificial intelligence to locate potential buyers in a matter of seconds, making it much more efficient than the massive listings used by traditional real estate agencies.
Instead of putting the information on a general sales listing that goes to everybody in the hopes that someone out there will be interested in the property, REX analyzes hundreds of thousands of consumer data points to identify specific people who might be interested in that specific house. For example, the company’s computers identify young couples who are buying baby clothes, because if they’re having a child, they might be needing a larger home. Or if someone is constantly buying construction materials and home repair products, they might be ready for a newer property. On the other hand, the REX algorithms will automatically ignore someone who just bought a massive flat-screen television, because there’s little chance that this person will be moving anytime soon. All this information is for sale, and it represents the daily bread of data analysts.
“The fact that you bought a hammer itself isn’t very informative to us,” says Jack Ryan, CEO of REX. “The changes in the rate at which you’re buying home supplies is very important to us.” Next, REX emails a first round of ads to thousands of potentially interested parties. Then if some five hundred people click on these ads in their email, the computer can find, in real time, the common characteristics among them, and email the ad to thousands of other people who share their profile.
The Robots Are Coming! Page 3