AI & The Future of Work 5
What technological change
means for the future of work
These developments in technology are exciting,
and bring enormous improvements, especially
for consumers and entrepreneurs, but they also
fuel fears that expanding articial intelligence
and machine capabilities may make humans
obsolete in the production process. Our ob-
jective is to discuss how and why these fears
come about, whether they were true in the
recent past, and to what extent they will apply
in the near future. Specically, we rst sketch a
framework that allows us to discuss the effects
of machines on employment. This framework
is an abstraction of the real world, but is useful
to assess how machines affect employment.
Second, we review how technology has affected
employment in the recent past. Third, we dis-
cuss what the future of work could look like.
Analytical framework
We adopt a framework developed by Acemoglu
and Autor (2011) on task allocation to discuss
how technology affects employment. The frame-
work starts with the premise that rms employ
workers to fulll tasks. A task is a unit of work
that is directly needed in production. Tasks differ
in terms of complexity. Workers supply effort to
perform tasks and differ in terms of their skills,
where “skill” refers to the capability to perform
tasks. Simple tasks can be performed easily by
skilled and unskilled workers, but complex tasks
Big data and advances in computing power have triggered a technological
revolution that may have enormous bearing on the workplace and the
labor market. Machines and robots are improving their capacities rapidly
through artificial intelligence (AI) and innovations in design and structure.
Digital assistants organize schedules, plan trips, and provide answers
to many questions people have. Autonomous cars drive around on our
streets and can bring customers from the pick-up point to any place
they desire.
Rafael Lalive and Daniel Oesch
University of Lausanne
can only be performed easily by skilled workers.
Workers earn wages that are in line with their
skills. In this context, rms will allocate tasks in
a very intuitive fashion. The low-skilled workers
perform the least complex tasks because their
comparative advantage, the ratio of output to
cost, is highest in these tasks. The high-skilled
workers will perform the most complex tasks,
and intermediate-skilled workers perform the
skills of intermediate complexity. Workers
perform the tasks that correspond to their
level of skill.
Technology in the form of machines, robots or
digital assistants competes with humans for
tasks. Machines in factories and computers in
workplaces have taken on the repetitive, but
cognitively demanding work of, for instance,
ofce clerks (automatic teller machines). Fewer
workers with intermediate skills are needed to
execute tasks of intermediate complexity, and
these workers then compete with both low-
skilled and high-skilled workers for low- and
high-complexity tasks. Intermediate-level jobs
will fare less well, with lower employment and
lower wages. According to this line of reasoning,
technology has led to a hollowing out of the
middle class, a phenomenon called polarization.
However, technology may also have been used
to replace workers of low complexity. In this
case, employment in highly-paid jobs grows
