Rise of the Robots: Technology and the Threat of a Jobless Future (5 page)

The Kroger Company, one of the largest grocery retailers in the United States, has also introduced highly automated distribution centers. Kroger’s system is capable of receiving pallets containing large supplies of a single product from vendors and then disassembling them and creating new pallets containing a variety of different products that are ready to ship to stores. It is also able to organize the way that products are stacked on the mixed pallets in order to optimize the stocking of shelves once they arrive at stores. The automated warehouses completely eliminate the need for human intervention, except for loading and unloading the pallets onto trucks.
²⁵
The obvious impact that these automated systems have on jobs has not been lost on organized labor, and the Teamsters Union has repeatedly clashed with Kroger, as well as other grocery retailers, over their introduction. Both the Kiva robots and Kroger’s automated system do leave some jobs for people, and these are primarily in areas, such as packing a mixture of items for final shipment to customers, that require visual recognition and dexterity. Of course, these are the very areas in which innovations like Industrial Perception’s box-moving robots are rapidly advancing the technical frontier.

The second transformative force is likely to be the explosive growth of the fully automated self-service retail sector—or, in other words, intelligent vending machines and kiosks. One study projects that the value of products and services vended in this market will grow from about $740 billion in 2010 to more than $1.1 trillion by 2015.
²⁶
Vending machines have progressed far beyond dispensing sodas, snacks, and lousy instant coffee, and sophisticated machines that sell consumer electronics products like Apple’s iPod and iPad are
now common in airports and upscale hotels. AVT, Inc., one of the leading manufacturers of automated retail machines, claims that it can design a custom self-service solution for virtually any product. Vending machines make it possible to dramatically reduce three of the most significant costs incurred in the retail business: real estate, labor, and theft by customers and employees. In addition to providing 24-hour service, many of the machines include video screens and are able to offer targeted point-of-sale advertising that’s geared toward enticing customers to purchase related products in much the same way that a human sales clerk might do. They can also collect customer email addresses and send receipts. In essence, the machines offer many of the advantages of online ordering, with the added benefit of instant delivery.

While the proliferation of vending machines and kiosks is certain to eliminate traditional retail sales jobs, these machines will also, of course, create jobs in areas like maintenance, restocking, and repair. The number of those new jobs, however, is likely to be more limited than you might expect. The latest-generation machines are directly connected to the Internet and provide a continuous stream of sales and diagnostic data; they are also specifically designed to minimize the labor costs associated with their operation.

In 2010, David Dunning was the regional operations supervisor responsible for overseeing the maintenance and restocking of 189 Redbox movie rental kiosks in the Chicago area.
²⁷
Redbox has over 42,000 kiosks in the United States and Canada, typically located at convenience stores and supermarkets, and rents about 2 million videos per day.
²⁸
Dunning managed the Chicago-area kiosks with a staff of just seven. Restocking the machines is highly automated; in fact, the most labor-intensive aspect of the job is swapping the translucent movie advertisements displayed on the kiosk—a process that typically takes less than two minutes for each machine. Dunning and his staff divide their time between the warehouse, where new movies arrive, and their cars and homes, where they are able to access and
manage the machines via the Internet. The kiosks are designed from the ground up for remote maintenance. For example, if a machine jams it will report this immediately, and a technician can log in with his or her laptop computer, jiggle the mechanism, and fix the problem without the need to visit the site. New movies are typically released on Tuesdays, but the machines can be restocked at any time prior to that; the kiosk will automatically make the movies available for rental at the right time. That allows technicians to schedule restocking visits to avoid traffic.

While the jobs that Dunning and his staff have are certainly interesting and desirable, in number they are a fraction of what a traditional retail chain would create. The now-defunct Blockbuster, for example, once had dozens of stores in greater Chicago, each employing its own sales staff.
²⁹
At its peak, Blockbuster had a total of about 9,000 stores and 60,000 employees. That works out to about seven jobs per store—roughly the same number that Redbox employed in the entire region serviced by Dunning’s team.

The third major force likely to disrupt employment in the retail sector will be the introduction of increased automation and robotics into stores as brick and mortar retailers strive to remain competitive. The same innovations that are enabling manufacturing robots to advance the frontier in areas like physical dexterity and visual recognition will eventually allow retail automation to begin moving from warehouses into more challenging and varied environments like stocking shelves in stores. In fact, as far back as 2005, Walmart was already investigating the possibility of using robots that rove store aisles at night and automatically scan barcodes in order to track product inventories.
³⁰

At the same time, self-service checkout aisles and in-store information kiosks are sure to become easier to use, as well as more common. Mobile devices will also become an ever more important self-service tool. Future shoppers will rely more and more on their phones as a way to shop, pay, and get help and information about
products while in traditional retail settings. The mobile disruption of retail is already under way. Walmart, for example, is testing an experimental program that allows shoppers to scan barcodes and then checkout and pay with their phones—completely avoiding long checkout lines.
³¹
Silvercar, a start-up rental car company, offers the capability to reserve and pick up a car without ever having to interact with a rental clerk; the customer simply scans a barcode to unlock the car and then drives away.
³²
As natural language technology like Apple’s Siri or even more powerful systems like IBM’s Watson continue to advance and become more affordable, it’s easy to imagine shoppers soon being able to ask their mobile devices for assistance in much the same way they might ask a store employee. The difference, of course, is that the customer will never have to wait for or hunt down the employee; the virtual assistant will always be instantly available and will rarely, if ever, give an inaccurate answer.

While many retailers may choose to bring automation into traditional retail configurations, others may instead elect to entirely redesign stores—perhaps, in essence, turning them into scaled-up vending machines. Stores of this type might consist of an automated warehouse with an attached showroom where customers could examine product samples and place orders. Orders might then be delivered directly to customers, or perhaps even loaded robotically into vehicles. Regardless of the specific technological path ultimately followed by the retail industry, it’s difficult to imagine that the eventual result won’t be more robots and machines—and significantly fewer jobs for people.

Cloud Robotics

One of the most important propellants of the robot revolution may turn out to be “cloud robotics”—or the migration of much of the intelligence that animates mobile robots into powerful, centralized computing hubs. Cloud robotics has been enabled by the dramatic
acceleration in the rate at which data can be communicated; it is now possible to offload much of the computation required by advanced robotics into huge data centers while also giving individual robots access to network-wide resources. That, of course, makes it possible to build less expensive robots, since less onboard computational power and memory are required, and also allows for instant software upgrades across multiple machines. If one robot employs centralized machine intelligence to learn and adapt to its environment, then that newly acquired knowledge could become instantly available to any other machines accessing the system—making it easy to scale machine learning across large numbers of robots. Google announced support for cloud robotics in 2011 and provides an interface that allows robots to take advantage of all the services designed for Android devices.
*

The impact of cloud robotics may be most dramatic in areas like visual recognition that require access to vast databases as well as powerful computational capability. Consider, for example, the enormous technical challenge involved in building a robot capable of performing a variety of housekeeping chores. A robotic maid tasked with clearing up the clutter in a room would need to be able to recognize an almost unlimited number of objects and then decide what to do with them. Each of those items might come in a variety of styles, be oriented in different ways, and perhaps even be somehow entangled with other objects. Compare that challenge to the one taken on by the Industrial Perception box-moving robot we met at the beginning of this chapter. While that robot’s ability to discern and grasp individual boxes even when they are stacked in a careless way is an impressive achievement, it is still limited to, well, boxes. That’s obviously a very long way from being able to recognize and manipulate virtually any object of any shape and in any configuration.

Building such comprehensive visual perception and recognition into an affordable robot poses a daunting challenge. Yet, cloud robotics offers at least a glimpse of the path that may eventually lead to a solution. Google introduced its “Goggles” feature for camera-equipped mobile devices in 2010 and has significantly improved the technology since then. This feature allows you to take a photo of things like landmark buildings, books, works of art, and commercial products and then have the system automatically recognize and retrieve information relevant to the photo. While building the ability to recognize nearly any object into a robot’s onboard system would be extraordinarily difficult and expensive, it’s fairly easy to imagine robots of the future recognizing the objects in their environment by accessing a vast centralized database of images similar to the one used by the Goggles system. The cloud-based image library could be updated continuously, and any robots with access to the system would get an instant upgrade to their visual recognition capability.

Cloud robotics is sure to be a significant driver of progress in building more capable robots, but it also raises important concerns, especially in the area of security. Aside from its uncomfortable similarity to “Skynet,” the controlling machine intelligence in the
Terminator
movies starring Arnold Schwarzenegger, there is the much more practical and immediate issue of susceptibility to hacking or cyber attack. This will be an especially significant concern if cloud robotics someday takes on an important role in our transportation infrastructure. For example, if automated trucks and trains eventually move food and other critical supplies under centralized control, such a system might create extreme vulnerabilities. There is already great concern about the vulnerability of industrial machinery, and of vital infrastructure like the electrical grid, to cyber attack. That vulnerability was demonstrated by the Stuxnet worm that was created by the US and Israeli governments in 2010 to attack the centrifuges used in Iran’s nuclear program. If, someday, important infrastructure
components are dependent on centralized machine intelligence, those concerns could be raised to an entirely new level.

Robots in Agriculture

Of all the employment sectors that make up the US economy, agriculture stands out as the one that has already undergone the most dramatic transformation as a direct result of technological progress. Most of those new technologies were, of course, mechanical in nature and came long before the advent of advanced information technology. In the late nineteenth century, nearly half of all US workers were employed on farms; by 2000 that fraction had fallen below 2 percent. For crops like wheat, corn, and cotton that can be planted, maintained, and harvested mechanically, the human labor required per bushel of output is now nearly negligible in advanced countries. Many aspects of raising and managing livestock are also mechanized. For example, robotic milking systems are in common use on dairy farms, and in the United States, chickens are grown to standardized sizes so as to make them compatible with automated slaughtering and processing.

The remaining labor-intensive areas of agriculture are primarily geared toward picking delicate, high-value fruits and vegetables, as well as ornamental plants and flowers. As with other relatively routine, manual occupations, these jobs have so far been protected from mechanization primarily because they are highly dependent on visual perception and dexterity. Fruits and vegetables are easily damaged and often need to be selected based on color or softness. For a machine, visual recognition is a significant challenge: lighting conditions can be highly variable, and individual fruits can be in a variety of orientations and may be partly or even completely obscured by leaves.

The same innovations that are advancing the robotics frontier in factory and warehouse settings are finally making many of these remaining agricultural jobs susceptible to automation. Vision
Robotics, a company based in San Diego, California, is developing an octopus-like orange harvesting machine. The robot will use three-dimensional machine vision to make a computer model of an entire orange tree and then store the location of each fruit. That information will then be passed on to the machine’s eight robotic arms, which will rapidly harvest the oranges.
³³
Boston-area start-up Harvest Automation is initially focused on building robots to automate operations in nurseries and greenhouses; the company estimates that manual labor accounts for over 30 percent of the cost of growing ornamental plants. In the longer run, the company believes that its robots will be able to perform up to 40 percent of the manual agricultural labor now required in the United States and Europe.
³⁴
Experimental robots are already pruning grapevines in France using machine vision technology combined with algorithms that decide which stems should be cut.
³⁵
In Japan, a new machine is able to select ripe strawberries based on subtle color variations and then pick a strawberry every eight seconds—working continuously and doing most of the work at night.
³⁶