Instead of asking, “What will the future of work look like?” I believe we
should ask, “What should the future
of work look like?” Instead of letting
ad-hoc, efficiency-centered decisions
drive the future, we need to make informed decisions with careful consideration of their organizational and societal impact. I am truly excited about
the future these algorithmic technologies can enable. I believe they can
usher in more efficient and fair management practices based on the best
of both data and human judgment.
They may create workplaces where
power structures are more equally
balanced between workers and managers thanks to the transparency of
their decisions. They may even enable
workers to make many of these managerial decisions themselves. What the
future of work will look like is up to
all of us who are living in this critical
[ 1] Lee, M. K., Kusbit, D., Metsky, E., and Dabbish, L.
Working with machines: The impact of algorithmic
and data-driven management on human workers. In
Proceedings of the 33rd Annual ACM Conference on
Human Factors in Computing Systems (CHI). ACM,
[ 2] Lee, M. K. and Baykal, S. Algorithmic mediation in group
decisions: Fairness perceptions of algorithmically
mediated vs discussion based social division.
Publication pending for Proceedings of the 20th ACM
Conference on Computer-Supported Cooperative Work
& Social Computing (CSC W). ACM, 2017.
[ 3] Lee, M. K., Forlizzi, J., Kiesler, S., Rybski, P., Antanitis,
J., and Savetsila, S. Personalization in HRI: A
longitudinal field experiment. In Proceedings of 7th
ACM/Institute of Electrical and Electronic Engineers
(IEEE) International Conference on Human-Robot
Interaction (HRI). IEEE, 2012, 319-326.
[ 4] Lee, M. K., Kiesler, S., Forlizzi, J., and Rybski, P.Ripple
effects of an embedded social agent: a field study of
a social robot in the workplace. In Proceedings of the
SIGCHI Conference on Human Factors in Computing
Systems (CHI). ACM, 2012, 695-704.
Min Kyung Lee is a research scientist in human-computer
interaction at the Center for Machine Learning and Health
and the Machine Learning Department at Carnegie Mellon
University. Her research examines the social and decision-making implications of intelligent systems and supports
the development of more human-centered machine
learning applications. Dr. Lee is a Siebel Scholar and has
received several best paper awards, as well as an Allen
Newell Award for Research Excellence. Her work has been
featured in media outlets such as the Ne w York Times,
New Scientist, and CBS. She received a Ph.D. in HCI and an
M. Des. in interaction design from Carnegie Mellon, and a
B. S. summa cum laude in industrial design from KAIS T.
© 2016 Copyright held by Owner(s)/Author(s).
Publication rights licensed to ACM.
assigned, other than to refuse the assignment. But many drivers created
workarounds by strategically controlling when and where they turned on
driver mode on the app in order to get
the types of requests and clientele they
wanted. They turned off driver mode
in bad neighborhoods to avoid dangerous situations, or went downtown
for successive short rides during the
In some cases, though, these
workarounds still don’t allow drivers enough control over their work.
For example, a former taxi driver now
working for Uber mentioned the lack
of choice in assignments made it hard
for him to create a work strategy. He
did not like the Uber assignment system because algorithms made the decisions that he used to make himself,
making him feel like he’d lost the
agency to enact strategies he’d developed to maximize his income. This
could be interpreted as resistance to
change, but also raises open-ended
ethics questions about the trend in
new technology to sacrifice individual control for the sake of overall system efficiency.
ROBOTIC COWORKERS THAT HELP
PEOPLE WORK TOGE THER
Another important aspect of successful workplaces is organizational culture. Algorithmic technologies in the
form of robots or virtual agents are increasingly taking on the role of cowork-er, either working directly with people
or sharing their physical workspace.
The question then becomes: How can
we make these new coworkers enjoyable to have around us? Can we leverage them to create better social and
organizational culture? For example,
how can they help workers socialize
better and increase communication
within and across teams?
To explore this question, we built
a delivery service robot from scratch
at Carnegie Mellon and deployed the
robot in an office building for a few
months in 2012 [ 3, 4]. Office employees
ordered apples, chocolate chip cookies, and other snacks online, and the
robot delivered them to their offices.
The robot called workers, initiated
brief small chats, and then asked them
to pick up their snacks.
The results of the field experiment
suggest workers built rapport with
the robot over time. What was more
surprising was the ripple effect that
it had on the social dynamics in the
workplace. The robot became a common boundary object that participants could easily relate to, creating a
topic of conversation and an occasion
to socialize, in the way that dogs do in
a public park. For example, one person
said: “It’s usually […] quiet in my hall.
You know, even if people are in, they
might close their door or something.
But I think people [were] more likely
to be around and laughing and feeling
sociable when the robot was there.”
This suggests a robot in a workplace
can have a positive impact on organizational culture.
However, I am not arguing all robots should be social and chatty.
The fact that the robot was social
and personalized for different individuals also had unintended consequences. Workers started to compare
each other. For example, one person
noted she felt jealous when the robot
complimented another girl. Interestingly, other people thought the robot
was nicer to this participant or even
flirting with her. This social comparison may have encouraged people
to use the robot more, be nicer to it,
and build stronger rapport with it.
However, it also created social tensions. Further research would need
to investigate how social and how
culturally aware a robot should be
for different contexts and purposes.
By taking advantage
of the pattern
people can analyze
from a thousand
to a million cases