In March of this year, a Google
computer defeated the world’s reigning Go champion in four out of five matches.
Several years earlier, IBM’s Watson computer defeated two of Jeopardy’s
greatest champions and IBM’s Deep Blue competed successfully against he former
world chess champion Garry Kasparov. We are all aware of the use of robots in
hospitals and combat zones, and the evolution of autonomous cars.
The military uses robots for several purposes,
including the recovery of improvised explosive devices. Hospitals are also
increasingly using robots for guiding patients and delivering drugs. More
recently, as reported in the Economist (2016), researchers have created “robodocs,”
robot surgeons that successfully stitched up the intestines of piglets with a
minimum of human supervision. Under a surgeon’s supervision, the Smart Tissues
Autonomous Robot (STAR) was able to sew piglets’ guts together after the doctors
had severed the piglets’ intestines. In fact, STAR was able to carry out about
60% of the procedure without human intervention and its stitches were more
evenly spaced and the sutured guts less leaky than what surgeons would have
done.
There is no question that we
are in the midst of a new age of artificial intelligence and the use of robots.
Bryjolffon and McAffee (2014) have argued that we are now in the second machine
age. The first started with James Watt’s steam engine, which kicked off the Industrial
Revolution. The second machine age started with using computers and digital
tools. Carlopio (1988) has described the different phases of new technology a
bit differently. In the 15th and 16th centuries,
printing, silk-throwing machinery, the screw press and the windmill were seen
as labor-saving devices. Traditional trades were not affected but were actually
enhanced. Then the Industrial Revolution began and technology became
labor-enslaving, with work processes becoming more standardized and specialized.
In the early twentieth century, scientific management became popular, with the
assembly line perhaps its most widely used application. Now we are in the third
phase of technology which he has described as labor-replacing, where computers
and robots will replace jobs.
There has been a lot written on
the impact of automation on workers and jobs, and I am not an expert in this
area. However, the consensus seems to be that computers are getting smarter.
Many experts predict that computers will displace jobs, not just at low-end but
also at the high-end. Acemoglu and Autor (2010) suggest that work can be
divided into a two-by-two matrix: cognitive versus manual and routine versus
nonroutine. Demand for routine tasks to be performed by humans has been falling
due to automation, whether these are routine cognitive tasks (e.g., bank
tellers, mail clerks) or routine manual tasks (e.g., machine operators, cement
masons, dressmakers). But nonroutine cognitive and manual work has been
growing, and there has been much debate on the extent to which computers and
robots will be able to perform these tasks and replace humans.
In the meantime, we are seeing more and more
robots working alongside humans. In 2012, Amazon bought Kiva Systems, a company
that makes robots. These robots are used in Amazon’s warehouses to “pick” and
bring goods from storage shelves to employees. What do we know about
interactions between humans and robots in the work setting? Actually, research
on human-robot interaction has been going on for a while. For example, Hinds et
al. (2004) wrote about the rise of “professional service” robots (as
distinguished from industrial robots) that share the workplace and help people perform
their tasks, e.g., supplying troops with ammunition in the battlefield,
delivering medications from pharmacies to nursing stations in hospitals. In the
future, these robots will be more highly interactive with people. The
researchers did a study to determine the effects of the robot’s appearance and
its relative status on how people work with robots. They created a lab
experiment that required subjects to interact with robots to accomplish some
tasks. They manipulated three levels of appearance (human as the baseline;
human-like where the robot had a face, torso, arms and legs, and wore an
outfit; and machine-like, where the robot covering was metallic and angular). Then
they manipulated status by telling the subjects that their robot partner was
their supervisor, their peer, or their subordinate. They found that subjects interacting
with a more machine-like robot had an increased sense of personal responsibility
they felt for the task, with subjects feeling most responsible when interacting
with a machine-like subordinate. Furthermore, subjects felt less responsible
when collaborating with a robot supervisor as compared with a robot peer or
subordinate.
As another example, Kim et al.
(2014) reviewed the literature on social distance and developed some hypotheses
on people’s reactions to interactions with robots. They had participants play a
card-matching game on a computer with Wakamaru, a robot developed by Mitsubishi
Heavy Industries, Ltd. As the researchers described it, here is what the robot
did:
“In its interactions with the participants, the
robot used three key behaviors: gaze, speech, and navigation. The robot’s
three-degrees-of-freedom head allowed it to direct its gaze toward the
participant and other targets in the environment. The robot communicated with
the participants using synthesized natural language and moved toward and away
from them at different points during the interaction.” (p. 786)
The participants played the
game with the robot, which made suggestions on moves. The researchers
manipulated power distance (with the robot either as supervisor or subordinate)
and proxemic distance (how close or distant the robot was to the
participant). What they found was that
participants who interacted with the supervisor robot at close distances
performed better and reported a more positive experience and stronger rapport
with the robot than those who interacted with the supervisor robot at far
distances. They also found that participants who interacted with a close
subordinate robot reported a more positive user experience and more rapport than
those who interacted with a distant robot.
What these and other studies
show is that there are interaction dynamics between humans and robots that we
need to take into account when designing the work of the future. Now if robots
can replace humans in many tasks, can they also replace humans as leaders? Here,
there is very little research on this topic. I did find an article by Samani et
al. (2012) who used the term robotics leadership to describe the work that robots
can perform in stock brokering (robots handling stock trades) and avionics
(robots replacing human pilots in airplanes). However, I don’t consider these as
leader behaviors as much as expertise that robots can provide. There is also some
emerging research on leader-like behaviors of robots among themselves; for
example, using simulation techniques among foraging robots, Pugliese et al.
(2015) found that the most skilled robots became “leaders” and that robot
groups with leaders were more effective than robot groups without leaders.
Part of being a leader is
influencing others to perform certain actions. We have seen evidence from
research by Milgram (1963) and others on the effect that those perceived to be
in positions of authority can have on getting others to comply and follow
orders. Parasuraman et al. (1997) wrote about this in the context of
automation, when they summarized the research on the dangers of automation. A
robot leader who gives instructions (especially if the robot has a deep male
voice and perhaps is made up to look dominant) might have others following
orders unquestioningly. In fact, in an interesting experiment where
participants worked with robots on a set of tasks, Gombolay et al. (2015) found
that participants preferred robots who made decisions about how the tasks were
to be allocated rather than the participants having to make the decisions
themselves. In another experiment, Robinette et al. (2015) found that
participants followed the robot in an artificially created emergency situation
(where they had to be evacuated from a room) even when they had seen the robot
make navigation mistakes earlier and continued to make mistakes in directing
them to a wrong exit!
We do know quite a bit about
effective leadership, and what people look for in leaders. Kousez and Posner (2007)
have been conducting surveys over the past 30 years on qualities most admired
in leaders, and they have found that four qualities are what most of the people
they surveyed around the world want in a leader: honest, forward-looing,
inspiring, and competent.
In his
now classic article, Kotter (2001) stated that leaders do things differently
than managers: leaders set a direction, align people, and motivate them. More
recently, Google found in its own research (Garvin, 2013) that outstanding
managers (versus average managers): coach well, empower their teams and do not
micro-manage, express interest in employees’ success and well-being, are
productive and results-oriented, are good communicators and listen to their
teams, help their employees with career development, have a clear vision and
strategy for the team, and have key technical skills so they can help advise
the team.
Can robots be as good as humans
in performing these behaviors? Potentially, yes. I can envision robot leaders
being programmed, for example, to express interest in employees’ success, to
listen and provide a clear vision for the team. I can also envision robot
leaders being programmed to be honest, forward-looking, inspiring, and
competent.
However, effective leadership
is not only about performing these behaviors. As Kousez and Posner (2007) argue,
leadership is about establishing a relationship between leader and follower.
While certain robot leadership behaviors might lead to compliance among
followers, organizations also want to create high-performing cultures
characterized by motivation and commitment. Can robots get us to trust them so
that they inspire us to do our best and engage us? Can a robot ever cause our
brains to release oxytocin, which is a chemical that helps promote many kinds
of social behavior (Stix, 2014)? Based on the evidence, it seems that our
brains are wired differently when it comes to reacting to those who inspire,
engage, and motivate us versus those who simply get us to comply.
As Colvin (2015) has pointed
out, we are asking the wrong question if we are trying to figure out only what
computers cannot do that humans can: “Rather than ask what computers can’t do,
it’s much more useful to ask what people are compelled to do – those things
that a million years of evolution cause us to value and seek from other humans,
maybe for a good reason, maybe for no good reason, but it’s the way we are.”
(p. 53)
In my view, we we will continue
to want human leaders in the work setting especially in three (and very “human”)
areas where leadership is critical: making decisions on business and
people issues (specifically around strategic decisions, and on who to select
and promote), communicating those decisions and related issues, and inspiring
and motivating. The value-add of human leaders is evident in these areas,
where commitment rather than compliance is of critical importance to building a
high-performance organization.
For each of the above, note the
following continuing patterns:
· Despite
years of evidence that statistical methods of selecting job candidates are
superior to human methods, almost every firm that I know still wants to see a
candidate face-to-face before hiring him or her, especially for higher-end work
and/or professional positions.
· When
managers congratulate someone for a job well done, or let them know that they
have been promoted, they much prefer to do this face-to-face rather than
sending them an e-mail or handing them a piece of paper with the news. The same
is true for communicating bad news, such as when someone has to be let go. Evidence
suggests that people also would much prefer to hear this information
face-to-face from their leaders.
· When a
team needs to be inspired and motivated, the most effective managers engage in
face-to-face meetings and “high-touch” actions to lift spirits up and boost
morale.
Brynjolfsson, E. and
McAfee, A. (2014). The Second Machine
Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies.
New York: W. W. Norton.
Carlopio, James.
(1998). Implementation: Making Workplace Innovation and Technical
Change Happen. Synergy Books International.
Colvin, G. (2015). Humans
Are Underrated: What High Achievers Know That Brilliant Machines Never Will.
New York: Penguin.
The Economist (2016). Who Wields the Knife? May 17, p. 74.
Garvin, D. (2013). How
Google Sold Its Engineers on Management. Harvard
Business Review.
Gombolay, Matthew C., et al.
(2015). Decision-making authority, team efficiency and human worker
satisfaction in mixed human–robot teams. Autonomous Robots 39
(3): 293-312.
Hinds, P., Roberts, R.
and Jones, H. (2004). Whose Job Is It Anyway? A Study of Human-Robot
Interaction in a Collaborative Task. Human-Computer
Interaction, 19: 151-181.
Kim, Y. and Mutlu, B.
(2014). How Social Distance Shapes Human-Robot Interaction. International Journal of Human-Computer
Studies, 72: 783-795.
Kotter, J. (2001). What
Leaders Really Do. Harvard Business
Review.
Kousez, J. and Posner,
B. (2007). The Leadership Challenge
(Fourth Edition). New York: Wiley.
Milgram,
S. (1963). Behavioral study of obedience. The Journal of Abnormal and
Social Psychology, 67 (4), 371-378.
Parasuruman, R. and
Riley, V. (1997). Humans and Automation: Use, Misuse, Disuse, and Abuse. Human Factors, 39 (2): 230-253.
Pugliese, F. et al. (2015). Emergence of
Leaders in a Group of Autonomous Robots. PLoS ONE 10(9):
e0137234. doi:10.1371/journal.pone.0137234.
Robinette, P et
al. (2015). Overtrust of Robots in Emergency Evacuation Scenarios. http://www.cc.gatech.edu/~alanwags/pubs/Robinette-HRI-2016.pdf
Samani, H. et al.
(2012). Towards Robotic Leadership. In SS. Par et al., (Eds.), ACHRS Part II. Hedelberg: Springer, pp.
158-165.
Stix, G. (2014). Fact or
Fiction: Oxytocin Is the Love Hormone. Scientific
American, September 8: http://www.scientificamerican.com/article/fact-or-fiction-oxytocin-is-the-love-hormone/
No comments:
Post a Comment