48 / Rotman Management Spring 20 19
University of Chicago economist Austan Goolsbee has argued,
a universal basic income might reduce labour market participa-
tion among low-wage groups.
AI AND COMPETITION. Currently, the leading companies in AI are
large in terms of revenue, profits and especially market capital-
ization (share price multiplied by the number of shares outstand-
ing). This has led to an increase in antitrust scrutiny of the leading
technology firms from governments (particularly the European
Commission) and in the popular press (e.g. The Economist’s cov-
er story “How to Tame the New Titans”). Much of this antitrust
scrutiny focuses on the role of these firms as ‘platforms’. Here,
we will focus on the particular feature of AI that is most relevant
to antitrust: the role of data.
Google Chief Economist Hal Varian has emphasized that
data is a scarce resource that exhibits decreasing returns to scale
in a technical sense: Prediction accuracy theoretically increases
in the square root of the number of observations, suggesting a
concave relationship between the amount of data and its value
in improving predictions. Under this argument, as University of
Washington Economist Patrick Bajari has argued, AI is unlikely
to generate antitrust concerns.
A more dynamic perspective suggests that there may be
economies of scale in terms of the business value of data. In
particular, if a slight lead in quality leads to a large lead in the
number of consumers, it could create a circle in which a slight
lead in data allows a company to collect more and better data,
reinforcing that lead and generating dominance over time. Thus,
although there may be decreasing returns to scale in a technical
sense, there may be increasing returns to scale in the economic
value of data. In such situations, the diffusion of AI may suggest
an increase in the importance of antitrust enforcement over time.
Finally, a more speculative antitrust concern with respect
to AI has been put forward by Oxford competition law expert
Ariel Ezrachi and University of Tennesee antitrust expert Mau-
rice E. Stucke in their book Virtual Competition: The Promise and
Perils of the Algorithm-Driven Economy, which is that firms will
deploy algorithms programmed to learn to set prices in oligopo-
listic competition with other firms and their algorithms.
It is possible that competing AIs — if given the goal of maxi-
mizing long-term profits — would learn to tacitly collude. We
stress, however, that this remains speculative. Current ‘game
playing’ AIs still operate in environments with a limited action
space — unlike the broad space that would cover pricing, let alone
multi-product pricing. A more likely application is that antitrust
authorities might be able to deploy AI techniques to identify
collusive behaviour.In closing
The research summarized herein makes three things clear: As
AI diffuses throughout our economy, it will have important con-
sequences for jobs, inequality and competition. Going forward,
both policy design and corporate strategy should focus on achiev-
ing the desired balance between encouraging diffusion without
compromising societal values.
While addressing these consequences will be the role of ed-
ucation policy, the social safety net and antitrust enforcement,
organizational leaders must carefully consider their own impact
as they make decisions that will have long-term consequences
for society.Ajay Agrawal is the Peter Munk Professor
of Entrepreneurship, Professor of Strategic
Management and Founder of the Creative
Destruction Lab at the Rotman School
of Management (and seven other locations).
Joshua Gans is Chief Economist of the Creative Destruction
Lab, the Jeffrey Skoll Chair of Innovation and Entrepreneur-
ship and Professor of Strategic Management at the Rotman
School. Avi Goldfarb is Chief Data Scientist
at the Creative Destruction Lab and the Ellison Professor
of Marketing at the Rotman School. They are the co-authors of Prediction
Machines: The Simple Economics of Artificial Intelligence (Harvard Business
Review Press, 2018). This article has been adapted from their paper,
“Economic Policy for Artificial Intelligence”, which was published in
Innovation Policy and the Economy (University of Chicago Press, 2019).Rotman faculty research is ranked #16 globally by the Financial Times.AI is likely to disproportionately increase
the wages of highly educated people.