TECHNOLOGY

What Is a Robot, Anyway?by H. James Wilson

APRIL 15, 2015

Quick — think of a robot, any robot.

If you’re into movies, you might imagine the android from the new movie Chappie, a machine

with artificial intelligence that learns morally questionable behavior. If you think more about

industry, you might imagine a mechanical arm programmed to install parts on a production

line. If you have dirty floors, you might think of a Roomba.

What do these and other robots have in common? What is it about them that makes them a

robot? Finding an all-encompassing definition of a robot is actually a difficult problem, even

for world-class roboticists. Form-factors, intelligence, and the purpose of robots can all vary

significantly. And yet many of us think we know a robot when we see one. How is this so?

For Westerners at least, our working cultural definition owes a lot to robots in stories and film,

as well as real-life robots past and present. But it’s worth looking further for a more

considered definition, starting with the origin of the word itself.

Before robot meant what it does today, the word meant “forced labor” or “hard work.” The

robot was a central European system of serfdom — according to the Oxford English Dictionary

— abolished in the Austrian Empire in 1848.

Then, in 1920, a Czech writer named Karel Čapek wrote a play called Rossum’s Universal

Robots, coining a new meaning for the word. In R.U.R., as it’s known, Čapek’s robots were

mass produced workers assembled from artificially synthesized organic material. The play

featured the first robot uprising, and the genre of dystopian robot sci-fi was born.

Descendants include Terminator and Battlestar Galactica, among others.

While some research groups are working hard to make highly intelligent humanoid robots as

those depicted in R.U.R., most real-life robotics efforts are decidedly less dramatic. A good

place to find the staid, real-life state of robots is the International Federation of Robotics, or

IFR. Helpfully, the IFR categorizes today’s robots into two major categories: industrial robots

and service robots.

The IFR defines an industrial robot as “an automatically controlled, reprogrammable,

multipurpose manipulator programmable in three or more axes, which may be either fixed in

place or mobile for use in industrial automation applications.” For years industrial robots were

all that a real robot could be.

The first industrial robot was installed in a Swedish metalworks plant in 1959. It was a

jointed, actuated arm that weighed two tons. Controlled by a program on a magnetic drum,

the robot relied on hydraulic actuators to adjust its position over a set of pre-programmed

joint angles. It was precise, but not necessarily elegant.

By 1973, there were 3,000 industrial robots in operation. By 2003, there were 800,000. Today,

more than 1.3 million industrial robots are in use or available in various industries including

automotive, electronics, rubber and plastics, cosmetics, pharmaceutical, and food and

beverage. Their market value is $9.5 billion. (A more complete history can be found here.)

Industrial robots will always have a place in the economy, but the relative newcomers on the

scene are service robots. This category, according to the IFR, is populated by autonomous

machines that complete tasks outside industrial applications. This means service robots are

found in personal and professional settings: a telepresence robot at work, a robot in the

operating room, an educational robot helping students learn to write code, a research robot

exploring the ocean, a robot in space helping astronauts make repairs, and so on.

The industrial vs. service distinction is helpful because it defines robots based on their

relationship with people and work more than around any technical factor. Combining

industrial and service robots, we can generalize to offer a basic definition of a robot as an

artificially created system designed, built, and implemented to perform tasks or services for

people.

There are echoes of the original meaning of the word robot here, as robots are doing labor and

hard work. And while this labor has historically been physical, we might also consider that

robots needn’t have actuating limbs.

Much of work today is knowledge work, therefore the definition of a robot should extend

even to automated computer programs to include cognitive computing, which describes IT

systems that can sense, comprehend, and act. This includes pre-programmed Twitterbots on

the low end and IPSoft’s Amelia artificial intelligence system on the high-end. Across the

spectrum, though, robots perform rule-based work, and tend to be configurable with basic

features like authentication, security, auditing, logging, and exception handling.

But even this broad definition will have to evolve as robots progress. What should we expect

from machines built to be stronger and smarter than the people who made them? Will they

always be limited to doing work for people? It’s a question that conjures Čapek’s robot

uprising, and has prompted many essays invoking Mary Shelley’s Frankenstein, among other

texts of monsters and unintended technological consequences.

It’s with good reason engineers, scientists, writers, ethicists, and philosophers are considering

the ramifications of advanced robotics and artificial intelligence. Even if real robots are

unlikely to match their dystopian sci-fi counterparts anytime soon, they still disrupt

economic sectors and directly affect the way people live and work.

Some clues to our robot future could come from an emerging field called “wise computing,”

presented at the recent annual meeting for the American Association for the Advancement of

Science (AAAS) in San Jose, Calif. Wise computing as a research field originated in Japan,

where cultural attitudes toward robots are less fraught as in Western countries.

The movement’s aims are to investigate the ethical, legal, and social relationships between

humans and machines, to develop machines that can make decisions with a kind of

programmatic wisdom, and to help humans make wiser decisions themselves. At the AAAS

panel, researchers from industry and universities in Japan and the UK discussed the

possibilities and challenges of such an approach.

While it might not catch on in the western world anytime soon, wise computing offers an

intriguing modification to the western cultural definition of robots. Imagine: Future robots

could be built to include a kind of ethical clause that limits what they are allowed to do. As we

progress with these technical, social, and ethical challenges, our definition of what a robot is

will have to change too.

H. James Wilson is Managing Director, Information Technology and Business Research, Accenture Institute

for High Performance. He is author of numerous HBR articles on human-technology interaction, including,

“Wearables in the Workplace.” You can follow him on Twitter @hjameswilson .

This article is about TECHNOLOGY

 FOLLOW  THIS TOPIC

Comments

Leave a Comment

POS T

0 COMMENTS

POSTING GUIDELINES

We hope the conversations that take place on HBR.org will be energetic, constructive, and thought-provoking. To comment, readers must sign

in or register. And to ensure the quality of the discussion, our moderating team will review all comments and may edit them for clarity, length,

and relevance. Comments that are overly promotional, mean-spirited, or off-topic may be deleted per the moderators' judgment. All postings

become the property of Harvard Business Publishing.

 JOIN THE CONVERSATION