A Blog by Jonathan Low

 

Jan 25, 2021

100 Years On, Robots' Reality Exceeds Fictional Expectations

The word 'robot' first came into common use in 1921. 

Though expectations remain exceptional compared to reality, the big difference between then and now is software. JL

Christopher Mims reports in the Wall Street Journal:

In a century-long dialogue between inventors of fictional and actual robots, engineers have for the most part been forced to play catch-up, either realizing or subverting the vision of robots first expounded in books, movies and television. Now, the reality of robots is in some areas running ahead of fiction, even ahead of what those who study robots for a living are able to keep track of. “The main difference between automation today and what we had 50 or 60 years ago is that we added software.”

On Jan. 25, 1921, Karel Čapek’s play “R.U.R.”—short for “Rossum’s Universal Robots”—premiered in Prague. It was a sensation. Within two years it had been translated into 30 languages, including English, to which it introduced the word “robot.” Čapek’s vision of unwilling slaves of humanity destined to rise up and destroy their makers has shaped our view of both automation and ourselves ever since.

In a century-long dialogue between inventors of fictional and actual robots, engineers have for the most part been forced to play catch-up, either realizing or subverting the vision of robots first expounded in books, movies and television.

Now, the reality of robots is in some areas running ahead of fiction, even ahead of what those who study robots for a living are able to keep track of.

Heather Knight is an engineer, “social roboticist” and one of 13 core faculty in Oregon State University’s robotics program. One day in late October, she was shocked to find the campus crawling with a fleet of autonomous, six-wheeled vehicles made by Starship Robotics. The San Francisco-based company had contracted with the campus dining service to provide contactless delivery.

“We’re at the point where not even the people in robotics know there are going to be robots on campus,” she adds.

This new visibility of robots—now in stores, hotels and health-care facilities, as well as on our streets and above our heads—is an indicator of their evolving nature. It’s also the outward sign of a watershed moment.

In 2019, 373,000 industrial robots were sold and put into use, according to the International Federation of Robotics, a not-for-profit industry organization that conducts an annual, global robot census based on vendor data. That number has grown about 11% a year since 2014, to a total of 2.7 million industrial robots in use world-wide. Industrial robots—descendants of the Unimate robot arm first installed at a General Motors factory in 1961—are the kind common in manufacturing, performing tasks like welding, painting and assembly. They work hard, but they’re not very smart. Also in 2019, 173,000 “professional service robots” were sold and installed, according to the federation. That number is projected to reach 537,000 units a year—a threefold increase—by 2023. These are the kind of robots businesses use outside of manufacturing. They perform a wide variety of functions, including defense, warehouse automation and disinfection in hospitals.

These robots tend to be much smarter, equipped with advanced software, sensors and Wi-Fi or other forms of connectivity. And instead of being hidden away in factories like industrial robots, they can generally do their jobs alongside people, instead of in walled-off areas where humans are forbidden to go.

If current growth rates for both of these kinds of robots hold, and we include professional-service robots not counted by the robotics federation, such as those made by companies solely for their own use, then some time in the coming year or so it’s likely service robots will overtake industrial robots in units sold or installed. That trend will bring new benefits for companies and consumers—and new challenges for workers.

By far the greatest share of professional service robots are those used in logistics. Mick Mountz, who founded Kiva Systems in 2003, helped pioneer the use of software, connectivity and sensors—coupled with off-the-shelf parts like motors, gearboxes, batteries and tires—to create relatively affordable robots that were more flexible and adaptable than their forebears, the industrial robots.

“The main difference between automation today and what we had 50 or 60 years ago is that we added software,” says Mr. Mountz. Just as critical was wireless connectivity—Wi-Fi was new at the time—and off-the-shelf sensors, like the black-and-white cameras used in the original Kiva robots, he adds.

Amazon bought Kiva in 2012, and it became Amazon Robotics. Kiva’s robots and software remain the foundation for perhaps the best-known example of service robots today: The rolling, pizza box-shaped “drive units” used by Amazon in its warehouses to move shelves of goods to the humans who pick and pack customers’ orders.

Many areas of our industrial world, from transportation to manufacturing, have yet to be affected by the newer kind of automation and robotics, but probably will be, adds Mr. Mountz.

The new generation of robots has already proved adaptable to an astonishing array of tasks, as evident in National Science Foundation-funded research conducted by Robin Murphy, director of the Humanitarian Robotics and AI Laboratory at Texas A&M University. Early in the course of the global pandemic, Dr. Murphy and her team set out to study the ways robots were being used to help humans adapt to the effects of Covid-19. She and her team documented 326 different robots, used in 29 different applications, telemedicine and hospital disinfection to quarantine enforcement, delivery, telepresence, construction, agriculture, logistics and laboratory automation.

Of these, 87% were existing robots adapted to help cope with the new virus, says Dr. Murphy. The sheer number and variety of mature robot technologies available for use in fighting the pandemic showed how companies and organizations are now spoiled for robot choice, she adds.

Oregon State’s Dr. Knight says the pandemic likely accelerated adoption of robots. While humans are generally averse to change, wars and natural disasters can inspire very rapid shifts. “There are different styles of innovation, but responding to necessity is one of the most impactful ones,” she says.

All the things that are not counted as robots by analysts and data-gatherers illustrate just how broad robotics has become, and how autonomy is now a thing that everyday objects can be imbued with. A full accounting could plausibly include all the autonomous drones recently authorized to fly in the U.S. by the Federal Aviation Administration, the many autonomous vehicles made by Waymo and its competitors, the rapidly growing population of smaller wheeled delivery robots, ocean-crossing autonomous ships, tens of millions of robot vacuum cleaners, a half-century of unmanned spacecraft, and, as they become ever more connected and “smarter,” perhaps even our homes.

Of course, more robots in the wild has meant more failures, too.

In 2019, the world’s first “robot hotel” was forced to eliminate more than half of its 243 robots, because they made life more difficult for guests and co-workers. In November 2020, Walmart scrapped its plans to have robots take inventory of its stores’ shelves, after discovering that humans could do the job more efficiently. Many more tasks for which robots seem well-suited have proved to be beyond their still quite-limited cognitive abilities, forcing companies to hire humans to pilot them remotely.

State of the (Robot) UnionThe world’s population of operationalindustrial robots continues to expand but at aslowing pace recently.Global number of industrial robots in use,change from a year earlierSource: International Federation of Robotics
%2010'15051015

Despite these teething problems, we are facing a likely acceleration in the pace at which robots and automation challenge American workers, says Mark Muro, a senior fellow at the Brookings Institution. Automation has always been most disruptive to the lives of blue-collar workers. But as robots transition from displacing people on the factory floor to displacing them in the service industry, which in the U.S. is many times larger than manufacturing, it could affect the lives of millions more people.

According to a historical analysis conducted by Mr. Muro and others at Brookings, recessions such as the one we’ve been in are the times businesses are most likely to replace humans with automation. This happens because, when recessions hit, revenue falls faster than wages. The result is that automation goes from a nice-to-have to a perceived necessity for cash-strapped companies.

Even when the economy recovers, that automation isn’t going away, adds Mr. Muro. While in the long run automation increases economic productivity and creates more jobs, in the short term it can mean unemployment and worse jobs for those swept aside by it. There is also the chronic and worsening problem that America is experiencing ever-greater economic inequality despite increased productivity from automation.


When she gave talks a decade ago, says Dr. Knight, she told her audiences that the robot revolution was already well under way, only it was happening behind closed doors, in places like factories and warehouses. What’s different now is that the robot revolution is happening in public, and is therefore unavoidable, even personal. In our homes, our places of work, on our streets, in our skies, robots are becoming a part of our everyday lives as they have never before.

A century after Čapek introduced the word robot to the English language, the one thing real-life robots have yet to do is run amok and destroy us all, as they did in his play and in countless works of science fiction since. But there’s one thing he did get right: As their ranks swell, and as they take on more tasks in more places, robots are, in their own way, taking over.

0 comments:

Post a Comment