by Ajay Agrawal
105
Biolo
Publications acr
2006
2015
2006
2015
2006
2015
–
–
wth
wth
wth
o
–
–
o
–
–
o
2004
2013
% gr
2004
2013
% gr
2004
2013
% gr
ning
able 4.6
obotics
T
Lear
systems
R
Symbol
systems
The Impact of Artifi cial Intelligence on Innovation 139
Table 4.7
Herfi ndahl- Hirschman index for application sectors
Application
H =
PatShare2
Chemical applications
153.09
Communications
140.87
Hardware and software
86.99
Computer science peripherals
296
Data and storage
366.71
Computer science business models
222
Medical applications
290.51
Electronic applications
114.64
Automotive applications
197.03
Mechanical applications
77.51
Other
129.20
This preliminary analysis does not trace out the important knowledge spill-
overs between innovation in the GPT and innovation and application sec-
tors, but it is probably far too early to look for evidence of this.
4.7 Deep Learning as a General Purpose Invention in the Method of
Invention: Considerations for Organizations, Institutions, and Policy
With these results in mind, we now consider the potential implications for
innovation and innovation policy if deep learning is indeed a general pur-
pose technology (GPT) and/or a general purpose invention in the method
of invention (IMI). If deep learning is merely a GPT, it is likely to generate
innovation across a range of applications (with potential for spillovers both
back to the learning GPT and also to other application sectors), but will not
itself change the nature of the innovation production function. If it is also
a general purpose IMI, we would expect it to have an even larger impact
on economy- wide innovation, growth, and productivity as dynamics play
out—and to trigger even more severe short- run disruptions of labor markets
and the internal structure of organizations.
Widespread use of deep learning as a research tool implies a shift toward
investigative approaches that use large data sets to generate predictions for
physical and logical events that have previously resisted systematic empirical
scrutiny. These data are likely to have three sources: prior knowledge (as in
the case of “learning” of prior literatures by IBM’s Watson), online transac-
tions (e.g., search or online purchasing behavior), and physical events (e.g.,
the output from various types of sensors or geolocation data). What would
this imply for the appropriate organization of innovation, the institutions
we have for training and conducting research over time, and for policy, par-
ticularly, as we think about private incentives to maintain proprietary data
sets and application- specifi c algorithms?
140 Iain M. Cockburn, Rebecca Henderson, and Scott Stern
4.7.1 The Management and Organization of Innovation
Perhaps most immediately, the rise of general purpose predictive ana-
lytics using large data sets seems likely to result in a substitution toward
capital and away from labor in the research production process. Many types
of R&D and innovation more generally are eff ectively problems of labor-
intensive search with high marginal cost per search (Evenson and Kislev
[1976], among others). The development of deep learning holds out the
promise of sharply reduced marginal search costs, inducing R&D organiza-
tions to substitute away from highly skilled labor toward fi xed cost invest-
ments in AI. These investments are likely to improve performance in existing
“search- intensive” research projects, as well as to open up new opportuni-
ties to investigate social and physical phenomena that have previously been
considered intractable or even as beyond the domain of systematic scientifi c
and empirical research.
It is possible that the ability to substitute away from specialized labor
and toward capital (that in principle could be rented or shared) may lower
the “barriers to entry” in certain scientifi c or research fi elds—particularly
those in which the necessary data and algorithms are freely available—while
erecting new barriers to entry in other areas (e.g., by restricting access to
data and algorithms). As of yet, there are few, if any, organized markets for
“trained” research tools or services based on deep learning, and few stan-
dards to evaluate alternatives. Our analysis suggests that the development
of markets for shared AI services and the widespread availability of relevant
data may be a necessary precursor to the broad adoption and dissemination
of deep learning.
At the same time, the arrival of this new research paradigm is likely to
require a signifi cant shift in the management of innovation itself. For ex-
ample, it is possible that the democratization of innovation will also be
accompanied by a lack of investment by individual researchers in special-
ized research skills and specialized expertise in any given area, reducing
the level of theoretical or technical depth in the workforce. This shift away
from career- oriented research trajectories toward the ability to derive new
fi ndings based on deep learning may undermine long- term incentives for
breakthrough research that can only be conducted by people who are at the
research frontier. There is also the possibility that the large- scale replace-
ment of skilled technical labor in the research sector by AI will “break
science” in some fi elds by disrupting the career ladders and labor markets
that support the relatively long periods of training and education required
in many scientifi c and technical occupations.
Finally, it is possible that deep learning will change the nature of scien-
tifi c and technical advance itself. Many fi elds of science and engineering are
driven by a mode of inquiry that focuses on identifying a relatively small
The Impact of Artifi cial Intelligence on Innovation 141
number of causal drivers of underlying phenomena built upon an under-
lying theory (the parsimony principle as restated by Einstein states that
theory should be “as simple as possible but no simpler.”) However, deep
learning off ers an alternative paradigm based on the ability to predict com-
plex multicausal phenomena using a “black box” approach that abstracts
away from underlying causes, but does allow for a singular prediction index
that can yield sharp insigh
t. De- emphasizing the understanding of causal
mechanisms and abstract relationships may come at a cost: many major
steps forward in science involve the ability to leverage an understanding
of “big picture” theoretical structure to make sense of, or recognize the
implications of, smaller discoveries. For example, it is easy to imagine a
deep learning system trained on a large amount of x-ray diff raction data
quickly “discovering” the double helix structure of DNA at very low mar-
ginal cost, but it would likely require human judgment and insight about a
much broader biological context to notice that the proposed structure sug-
gests a direct mechanism for heredity.
4.7.2 Innovation and Competition Policy and Institutions
A second area of impact, beyond the organization of individual research
projects or the nature of what counts as “science” in a particular fi eld, will
be on the appropriate design and governance of institutions governing the
innovation process. Three implications stand out.
First, as discussed earlier, research over the past two decades has empha-
sized the important role played by institutions that encourage cumulative
knowledge production through low- cost independent access to research
tools, materials, and data (Furman and Stern 2011; Murray and O’Mahony
2007). However, to date there has only been a modest level of attention to
the questions of transparency and replicability within the deep learning
community. Grassroots initiatives to encourage openness organized through
online hubs and communities support knowledge production. But it is useful
to emphasize that there is likely to be a signifi cant gap between the private
and social incentives to share and aggregate data—even among academic
researchers or private- sector research communities. One implication of this
divergence may be that to the degree any single research result depends on
the aggregation of data from many sources, it will be important to develop
rules of credit and attribution, as well as to develop mechanisms to replicate
the results.
This implies that it will be particularly important to pay attention to
the design and enforcement of formal intellectual property rights. On the
one hand, it will be important to think carefully about the laws that cur-
rently surround the ownership of data. Should the data about, for example,
my shopping and travel behavior belong to me or to the search engine or
ride- sharing company that I use? Might consumers have a strong collective
142 Iain M. Cockburn, Rebecca Henderson, and Scott Stern
interest in ensuring that these data (suitably blinded, of course) are in the
public domain so that many companies can use them in the pursuit of inno-
vation?
On the other hand, the advent of deep learning has signifi cant implica-
tions for the patent system. Though there has so far been relatively little
patenting of deep learning innovations, historical episodes such as the dis-
covery and attempted wholesale patenting of express sequence tags and
other kinds of genetic data suggests that breakthroughs in research tools—
often combined with a lack of capacity at patent offi
ces and confl icting court
decisions—can result in long periods of uncertainty that has hampered the
issuing of new patents, and this in turn has led to lower research productiv-
ity and less competition. Deep learning also presents diffi
cult questions of
legal doctrine for patent systems that have been built around the idea of
creative authors and inventors. For example, “inventorship” has a specifi c
meaning in patent law, with very important implications for ownership and
control of the claimed invention. Can an AI system be an inventor in the
sense envisaged by the drafters of the US Constitution? Similarly, standards
for determining the size of the inventive step required to obtain a patent
are driven by a determination of whether the claimed invention would or
would not be obvious to a “person having ordinary skill in the art.” Who
this “person” might be, and what constitutes “ordinary skill” in an age of
deep learning systems trained on proprietary data are questions well beyond
the scope of this chapter.
In addition to these traditional innovation policy questions, the pros-
pect for deep learning raises a wide variety of other issues, including issues
relating to privacy, the potential for bias (deep learning has been found to
reinforce stereotypes already present in society), and consumer protection
(related to areas such as search, advertising, and consumer targeting and
monitoring). The key is that, to the extent that deep learning is general
purpose, the issues that arise across each of these domains (and more) will
play out across a wide variety of sectors and contexts and at a global rather
than local level. Little analysis has been conducted that can help design
institutions that will be responsive at the level of application sectors that
also internalize the potential issues that may arise with the fact that deep
learning is likely to be a GPT.
Finally, the broad applicability of deep learning (and possibly robotics)
across many sectors is likely to engender a race within each sector to establish
a proprietary advantage that leverages these new approaches. As such, the
arrival of deep learning raises issues for competition policy. In each appli-
cation sector there is the possibility of fi rms that are able to establish an
advantage at an early stage, and in doing so position themselves to be able to
generate more data (about their technology, about customer behavior, about
their organizational processes), and will be able to erect a deep- learning-
driven barrier to entry that will ensure market dominance over at least the
The Impact of Artifi cial Intelligence on Innovation 143
medium term. This suggests that rules ensuring data accessibility are not
only a matter of research productivity or aggregation, but also speak to
the potential to guard against lock-in and anticompetitive conduct. At the
present moment there seem to be a large number of individual companies
attempting to take advantage of AI across a wide variety of domains (e.g.,
there are probably more than twenty fi rms engaging in signifi cant levels of
research in autonomous vehicles, and no fi rm has yet to show a decisive
advantage), but this high level of activity likely refl ects an expectation for
the prospects for signifi cant market power in the future. Ensuring that deep
learning does not enhance monopolization and increase barriers to entry
across a range of sectors will be a key topic going forward.
4.8 Concluding
Thoughts
The purpose of this exploratory chapter has not been to provide a sys-
tematic account or prediction of the likely impact of AI on innovation, nor
clear guidance for policy or the management of innovation. Instead, our
goal has been to raise a specifi c possibility—that deep learning represents a
new general purpose invention of a method of invention—and to draw out
some preliminary implicati
ons of that hypothesis for management, institu-
tions, and policy.
Our preliminary analysis highlights a few key ideas that have not been
central to the economics and policy discussion so far. First, at least from the
perspective of innovation, it is useful to distinguish between the signifi cant
and important advances in fi elds such as robotics from the potential of a
general purpose method of invention based on application of multilayered
neural networks to large amounts of digital data to be an invention in the
method of invention. Both the existing qualitative evidence and our pre-
liminary empirical analysis documents a striking shift since 2009 toward
deep learning- based application- oriented research that is consistent with
this possibility. Second, and relatedly, the prospect of a change in the innova-
tion process raises key issues for a range of policy and management areas,
ranging from how to evaluate this new type of science to the potential for
prediction methods to induce new barriers to entry across a wide range of
industries. Proactive analysis of the appropriate private and public policy
responses toward these breakthroughs seems like an extremely promising
area for future research.
144 Iain M. Cockburn, Rebecca Henderson, and Scott Stern
Appendix
Table 4A.1
Artifi cial intelligence keyword allocation
Symbols
Learning
Robotics
Natural language processing
Machine learning
Computer vision
Image grammars
Neural networks
Robot
Pattern recognition
Reinforcement learning
Robots
Image matching
Logic theorist
Robot systems
Symbolic reasoning
Bayesian belief networks
Robotics
Symbolic error analysis
Unsupervised learning
Robotic
Pattern analysis
Deep learning
Collaborative systems
Symbol processing
Knowledge representation and reasoning
Humanoid robotics
Physical symbol system
Crowdsourcing and human computation
Sensor network
Natural languages
Neuromorphic computing
Sensor networks
Pattern analysis