Digital Technologies, Innovation, and Skills: Emerging Trajectories and Challenges

Forthcoming in Research Policy

April 2021

ABSTRACT

In order to better understand the complex and dialectical relationships between digital technologies,

innovation, and skills, it is necessary to improve our understanding of the coevolution between the

trajectories of connected digital technologies, firm innovation routines, and skills formation. This is

critical as organizations recombine and adapt digital technologies; they require new skills to innovate,

learn, and adapt to evolving digital technologies, while digital technologies change the codification of

knowledge for productive and innovative activities. The coevolution between digital technologies,

innovation, and skills also requires, and is driven by, a reorganization of productive and innovation

processes, both within and between firms. We observe this in all economic sectors, from agriculture to

services. Based on evidence on past technologies in the innovation literature, we suggest that we might

require a new set of stylized facts to better map the main future trajectories of digital technologies, their

adoption, use, and recombination in organizations, to improve our understanding of their impact on

productivity, employment and inequality. The papers in this special issue contribute to a better

understanding of the interdependence between digital technologies, innovation, and skills.

1
1. Introduction

2
 Digital technologies permeate and restructure all facets of economic and social activities. In some

ways that they are implemented, they disrupt existing activities, while in others they have a more

incremental impact and complement existing activities. In some cases, they substitute for existing

technologies and tasks, but in others they are complementary. At times, they result in the creation of

new activities, services, innovation, and business opportunities. Digitalization is powerful because it

not only allows automation but also tracks and stores information and data about tasks and activities,

thereby creating a record that can be analyzed and that provides opportunities to improve processes,

work organization (Zuboff, 1988), and predictions about future events (Agrawal, Gans, and Goldfarb

2018). The ability to model the analog world digitally has unleashed a wave of both innovation and

hype. Widely discussed technologies include the internet of things, blockchains, additive

manufacturing, big data, artificial intelligence, cloud computing, and augmented and virtual reality (for

one of many listings, see Rindfleisch et al., 2017). Some of them, such as cloud computing, are already

realized, others may never become significant, and new combinations certainly will emerge.

3
 The application domains of digital technologies span all sectors, from agriculture and

manufacturing to professional services, health services, and beyond. Much popular and academic

attention has been paid to the transformational or disruptive implications of digital technologies for

businesses (new business models, types of products/services, types of customer experiences, and

organizational structures and routines), work, education, and society at large. Organizations and

workers might need to adapt or even radically transform themselves to succeed in the continually

evolving digital world (Brynjolfsson and McAfee, 2014; Frey and Osborne, 2017; Nambisan et al.,

2019; Schwab, 2017). Even changes that seem incremental and gradual and follow the learning

processes typically observed for new technologies might ultimately become deeply transformational in

ways that are unpredictable ex ante. For instance, the ability to codify human tasks in software has a

profound effect on skills and the types of jobs available (e.g., Helper et al., 2021). Skill requirements

are changing within and across organizations, industries, and countries and making existing ones

redundant or obsolete (Autor, 2015; Autor et al., 2015; Cedefop, 2018).

The time scales and directions of these changes are heterogeneous across technologies,

industries, and locations. Multiple technological trajectories, for different digital technologies and

applications, may develop, clash, and evolve over time, at a different pace in different industries

(Martinelli et al., 2021). Digital technologies also provide the means to reinvent firms, improving the

organization and production of goods and services as well creating new goods and services and even

transforming innovation routines (Alcacer et al., 2016; Hinings et al., 2018; Nambisan et al., 2017).

4
 The direction of change in the adoption and use of digital technologies at organizations and the

adoption of innovations in previous routines can also be affected by unexpected events. For instance,

the COVID-19 pandemic forced many in-person events to go online within a very short time. The use

of digitalization that previously had a limited impact, such as telemedicine, was massively increased in

a matter of weeks (e.g., Mann et al., 2020). Although the adoption of telemedicine implies a “simple”

shift from in-person to video consultation, for many regular doctors’ visits, the advantages (and

potential drawbacks) became obvious immediately, and the earlier routines might not be restored.

Changes that might have taken years to be adopted were accelerated by this unexpected event. 1

1
The shift is likely to unleash many changes and innovations concerning how to manage this transformation most

efficiently and effectively. There are also likely to be major implications for employment patterns, as there will be

significantly reduced demand for receptionists, janitors, and various supplies. A picture taken at home by a high-resolution

smartphone is becoming sufficient for preliminary diagnosis. New innovations, such as tele-diagnosis equipment could be

used.
5
 In order to better understand the complex and dialectical relationships between innovation,

skills, and digital technologies, it is necessary to improve our understanding of the coevolution between

the trajectories of connected digital technologies, firm innovation strategies, and skills. Advances in

digital technologies are not completely exogenous. Innovation is an outcome as well as a source of

digital transformations in the economy and society. Digital technologies require unique skills to be

developed as well as different skills to be introduced in production processes, to support new

organization of innovating firms, its buyers and suppliers, even as they result in a reconfiguration of the

division of labor. These factors also influence the innovation process and the development of new

digital technologies. As digital technologies continue to evolve, the skills needed to develop and adopt

these technologies also are evolving. In fact, the ability to adapt to this constant evolution suggests that

firms and their employees must develop a “meta-skill,” which is the ability to adapt existing skills

continually or acquire new skills to use the new capabilities afforded by change in the digital

technologies.

We illustrate the interdependence between innovation, skills, and digital technologies in Figure

1. The three vertices show the dyadic relationships between the three dynamics, which are usually

studied in pairs. The systemic interaction between these three dynamics requires and is driven by a

reorganization of productive and innovation processes, both within and between firms. Based on the

literature, in this paper, we first illustrate the interconnections between these three dynamics and then

discuss implications for firm and industry organization, including how these organizational changes

influence the three dynamics and their coevolution.

6
Figure 1. Interconnections between innovation, skills, and digital technologies

The remainder of this paper is organized as follows. Building on examples of specific digital

technologies, Section 2 explores their interdependence with innovation and skills. In Section 3 we

discuss the relationship between digital technologies, skills, and innovation and the reorganization of

activities. We discuss how these reorganizations differ across activities in the primary sector,

manufacturing, and services. Based on the questions raised by this discussion, Section 4 elaborates

three research challenges for gaining a better understanding of the coevolution between innovation,

skills, and digital technologies and the relevant impacts. In Section 5, we discuss the contributions

toward answering these questions made by the seven papers in this special issue. In Section 6, we offer

some implications and suggest a few directions for future research.

7
2. Exploring the Coevolution among Digital Technologies, Innovation, and Skills

The result of von Neumann–based digital computing is the proliferation of individual technologies,

ranging from computers and industrial robots, which have been in use for decades, to “intelligent”

machines that can “learn” (machine learning). Given the diversity of technologies, not surprisingly,

different studies provide different definitions and categorizations of digital technologies. These

technologies also differ with respect to the rate at which they have been adopted, by which firms, and

how they interact, although overall their impact and continuing momentum are enormous. For example,

robotics is a comparatively well-established technology in manufacturing, but, more recently, it has

been widely adopted in the services, in particular, retail and warehousing, and, even more broadly, in

all kinds of activities, such as answering phone calls (chatbots). The adoption of robotics in

manufacturing is highly skewed and mostly adopted by just a few large firms (Acemoglu, Le Large,

and Restrepo, 2020; Deng, Plümpe, and Stegmaier, 2020).

One vision of the future of digital technologies, the Industry 4.0 “paradigm,” was introduced by the

German government in 2013 and has attracted a great deal of attention (Kagermann, 2015; Lasi et al.,

2014). Industry 4.0 envisioned a “smart” factory based on cyberphysical systems that integrate

advanced technologies in the physical and digital production systems and allow integration across the

data systems of firms and sectors. It was intended to drive the integration of technologies, such as

factory automation/robots, data exchange in manufacturing technology, additive manufacturing/3D

printing, cloud computing, the internet of things, big data, and AI (e.g., Martinelli et al., 2021).

Whereas Industry 4.0 focused on manufacturing, other studies take a broader view of the diffusion of

digital technologies and recognize their application beyond manufacturing (e.g., Lusch and Nambisan,

2015).

8
 In Figure 2, we identify some of the most significant digital technologies that are reconfiguring the

workplace, when they initially emerged, and their data intensity. These technologies range from those

based on data, such as cloud computing and AI, to those with a physical manifestation, such as robots

and drones. The most data-intensive technologies in particular are progressing rapidly, based on the

dynamic of Moore’s law regarding the speed of improvements in processing power. Because their

underlying “raw material” is processed data, they can be combined and recombined in ways that enable

new applications, which can create new value (Henfridsson et al., 2018). This opens up a seemingly

infinite set of options for further recombinations and the initiation of various technological trajectories

(Lanzolla et al., 2021). For example, new fruit-picking robots combine robotics with image recognition

to determine which pieces of fruit to pick and which to leave for further maturation. Experimentation is

ongoing on drones that will not only pick the fruit but also deposit it in the crates for shipping. For

another example, drones are being trained to identify different kinds of weeds and spray them with

microbursts of pesticide (for a general discussion, see Kenney et al., 2020). These recombinations and

the development of task-specific machines and software make digital technologies generative of yet

more innovations (Zittrain, 2008).

9
 Figure 2. Digital Technologies: Emergence, Evolution, and Fusion

To distinguish current from previous waves of automation, Shoshona Zuboff (1988) labeled this

digitization process as one in which a machine is equipped with digital sensors. She suggested that by

digitizing its operations, its events and activities became transparent to the organization and thus more

easily optimized, linked to yet other machines, and reorganized—all according to a logic dictated by

computers. These characteristics allow the technology to coevolve with firm innovation and

organization in ways that enable the creation of new business models and value propositions (Colfer

and Baldwin 2016). As this process advances, new and, at times, unexpected recombinant innovations

emerge. This is illustrated in the box on the right-hand side of Figure 2 that shows how these

technologies are recombined and integrated, creating yet more innovation.

The interdependence of these digital technologies is confirmed by the adoption patterns.

Recent surveys show that although the rate of adoption for many of these digital technologies is

relatively low and skewed toward larger firms, it has a hierarchical pattern in which the most

10
sophisticated technologies are most frequently accepted only after more basic applications (e.g., Zolas

et al., 2021). Moreover, substantial heterogeneity is seen in the recombination and development of new

technologies that cannot be fully explained by adoption data. For example, many firms develop their

digital technologies in house for their own use (Montobbio et al., 2020). Early developers and adopters

appear to have competitive advantages, and research suggests that the gap between early adopters and

laggards is likely to grow (Barth et al., 2020a, 2020b). Presumably, innovation is already underway by

these early adopters to integrate and fully exploit the potential of the new technologies. This

interdependence between digital technologies and innovation processes makes it more difficult to

predict their development and when and where a transformative effect will take hold. Our

understanding of the evolution and impact of these digital technologies should improve over time as we

acquire more data (Frank et al., 2019).

Overall, it is not yet clear how these digital technologies, all of which are evolving rapidly, will

coevolve with the reorganization of tasks within occupations and across firms (Agrawal et al., 2019;

Lane and Saint-Martin, 2021). Much of our knowledge about the relationship between digital

technologies and skills relies on quantitative research on the adoption of information and

communication technology (ICT) (e.g., Autor et al., 1998; Bresnahan et al., 2002) and automation (e.g.,

Arntz et al., 2017; Bessen, 2019) and the extent to which they replace humans in performing tasks

(Acemoglu and Restrepo, 2018).

Skill-biased technological change (Autor et al., 1998; Bresnahan et al., 2002) and, subsequently,

routine-biased technological change (Goos et al., 2009, 2014) have been proposed to explain historical

patterns in the labor market, such as the relative decline in low-skill workers compared with both high-

skill and medium-skill workers at the beginning of the 1980s (Goos and Manning, 2007; Goos et al.,

2009; Adermon and Gustavsson, 2015), and the more recent hollowing out of the middle class and

11
middle-skilled routinized tasks, which are susceptible to being performed by machines (Autor, 2014;

2015). Based on this framework, estimates of the impact of digital technologies on employment vary

widely. Some views are more pessimistic, i.e., digital technologies will mainly lead to the elimination

of some jobs (Acemoglu and Restrepo, 2019; Frey and Osborne, 2017; Korinek and Stiglitz, 2017);

others are more optimistic, i.e., digital technologies will mainly improve existing jobs or create new

ones (Arntz, Gregory, and Zierahn, 2017; Autor and Salomons, 2018; Felten, Raj, and Seamans, 2019);

and still others take a more neutral stance, i.e., the effects will be mixed (Das et al., 2020; Nedelkoska

and Quintini, 2018). These studies tend to consider the technologies relatively well defined

(skill/routine biased) and exogenous, which is at odds with the coevolution with innovation processes

and reconfigurations that they seem to undergo within firms.

The sectors that have adopted digital technologies with the greatest enthusiasm include finance,

insurance, mass production, and process industries. Manufacturing is commonly believed to be at the

forefront of game-changing technological innovation, but far less attention has been paid to the

services. For example, a report on the impact of advanced industrial robotics, additive manufacturing,

the industrial internet of things, and electric vehicles suggests that these new technologies are likely to

affect employment by displacing workers, but it is hoped that they will also result in an upgrading of

occupations, the development of more hybrid skill sets, and a decrease in repetitive routine work

(Eurofound, 2018).

Over the past few decades, three main patterns of labor market change occurred as a result of these

changes in organization (Eurofound, 2020): (1) upgrading (a linear improvement in the employment

structure, with highest employment growth in highly paid jobs); (2) polarization (larger employment

growth at both ends of the job–wage distribution, shrinking the middle), and (3) flexibilization (an

increase in nonstandard employment, characterized by marginal part-time work, short-term temporary

12
contracts, and on-demand work, often called “gig” work (De Stefano, 2015).2 For each of these

patterns, digitalization appears to have played a role in increasing their prevalence. 3

The debate over routinized tasks, and what can be automated, harkens back to Polanyi's distinction

between “tacit” and “explicit” knowledge. However, the distinction between them has evolved as

digital technologies have improved. For instance, what used to be nonroutine and tacit has

progressively become standardized, formalized, better understood, and eventually replicated by

software and computers and thus has become routine and codified (i.e., explicit). The advances in

digital technologies in terms of computational power, availability, and handling of big data to train

machines, storage capacity, and machine learning algorithms enable algorithmic processing of what

previously were considered nonroutine tasks that required human learning. The occupations affected

range from creative and engineering design fields to legal and other professions (Susskind, 2017).

The key to the ability of these advanced digital technologies to undertake what previously were

considered nonroutine tasks is not to replicate human processes and thinking (Susskind and Susskind,

2016) but to standardize some parts of the overall task in order to make them amenable to

computerized processing. After nonroutine complex (i.e., integrated) processes are broken down in

modular components, they become easier to automate. Reducing the degree of complexity and

modularizing complex processes (Kauffman and Levin, 1987; Simon, 1969, 2002) enable machines to

replace or expand and complement human capabilities (O’Donovan and Smith, 2020).

2
There is ample evidence for this digitalization-enabled work, even in cases, such as Uber, at least some of the workers
appreciate the flexible arrangements, despite the low pay, precarity, and inability to support a family on the basis of the
income. Similarly, in IT-enabled gig work, such as that provided by Upwork; some number of the workers find the work
convenient and attractive. In more skilled tasks, it has been demonstrated that some workers choose to work on gigs,
rather than accept full-time employment (see, for example, Barley and Kunda 2011). Thus, while the work can be seen as
exploitative, even among the workers there appears to be no uniform sentiment.
3
We do not attribute all of the increased prevalence of these types of contractual arrangements solely to digitalization
and thus ignore the role of politics and the neoliberal ideology, see, for example, in the case of the US, Weil (2014).
13
 Alongside the new dynamics of substitution between machines and workers performing non-

routine cognitive tasks, new dynamics of complementarity among tasks are likely to emerge and

become more important, especially because creative tasks, a large subset of nonroutine cognitive tasks,

have become more important as the fourth industrial revolution unfolds (Schwab 2017; Pedota and

Piscitello, 2020). Machines can complement people in performing tasks that require processing massive

amounts of information and data, thus supporting the experiential and emotional judgment of

professionals, such as doctors and lawyers. For example, a medical diagnosis of skin cancer based on

dermatological images, which use the results of biopsies, can employ AI-based software with higher

diagnostic accuracy than trained physicians (Esteva et al., 2017), though computers and people make

different types of mistakes (Brynjolfsson and Mitchell, 2017).

Other activities are more difficult to decompose into codifiable and automatable tasks. For

instance, tasks that require the recognition of emotions and formulation of a response to them are more

difficult to codify in the current state of digitalization. If this is the case, then social skills should

increase in importance (e.g., Autor et al., 2002; Brynjolfsson and MacAfee, 2014; Deming, 2017). Paul

Deming (2017) suggests that possessing interpersonal and social skills has higher returns than STEM

skills.

The direction of progress in machine learning is crucial in this debate, because it will determine

which activities and tasks will be liable to replace people with technology (Brynjolfsson and Mitchell,

2017). AI algorithms are increasingly competent at performing specific tasks that have traditionally

required human expertise, with emerging applications in medicine, law, transportation, scientific

discovery, and other industries (Esteva et al., 2017; Wilder et al., 2020). As the technology coevolves

with firm innovation and skills, the extent to which they will replace/complement human occupations is

still unclear. Adopting a task-based perspective, Acemoglu et al. (2020) study AI adoption through

14
firms’ posting of vacancies that require workers who specialize in AI-related activities. They find that

firms that hire AI-proficient workers replace older skills with new skills, modifying their task structure

when they replace tasks that were performed by humans, and find no evidence of complementarity

between human and AI tasks (although they do not find any employment effect at the industry level).

Felten, Raj, and Seamans (2019) study progress in AI tasks (measured by the Electronic Frontier

Foundation [EFF] AI Progress Measurement dataset) in relation to the various abilities that characterize

the jobs listed in the O*NET (the US Department of Labor’s evolving list of occupations). They use

this relationship to study the effect of advances in specific AI technologies on different occupations

(Felten, Raj, and Seamans, 2019). They find that the occupations that are more affected by advances in

AI have increases in employment; the more this is the case, the more automated the occupations are.

They interpret their result as an increase in complementarity between AI and human skills, especially in

firms that were already on an automation trajectory, which is augmented by AI. By connecting AI

patent descriptions to the occupational descriptions in O*NET, Webb (2019) uses a slightly different

strategy to connect AI to tasks and finds that, in contrast to earlier rounds of automation related to

software and robots, AI may be able perform tasks that previously were considered highly skilled.

Building upon Figures 1 and 2, it is important to consider the convergence and fusion of characteristics

of digital technologies and skills, depicted in Figure 3. The horizontal axis shows the characteristics of

digital technologies, as a spectrum that goes from fully tangible or physical to completely intangible or

computational. On the vertical axis, we show the characteristics of skills ranging from STEM to soft

skills (including creative, interpersonal, and social skills), which are related to these technologies. No

technologies are in the top-left quadrant, combining physical technologies and soft skills; however, this

set of skills is gaining importance in the current generation of digital technologies. Figure 3 illustrates

the integration of early generations of digital technologies into the current generation of digital

15
technologies, spanning the full range of characteristics (from physical to intangible) and skills (from

STEM to soft).

Figure 3. The Convergence of Digital Technologies and Skill Characteristics

3. The Reorganization of Production and Innovation Processes

This coevolution of digital technologies, innovation, and skills affects the ways in which firms

organize activities within and between them (Hitt, 1999). Here we give some examples, distinguishing

between primary, manufacturing, and service activities.

Digitalized activities appear to be particularly placeless, in that the work can be done anywhere

with an internet connection and a computer. In the first decade of the 2000s, it appeared that countries

with low labor costs, such as India and the Philippines, would benefit the most from this digitally

enhanced flexibility (Kenney et al., 2009). In retrospect, the apocalyptic visions of offshoring work
16
have not yet been realized, perhaps because the cost differentials became less attractive over time and

because many firms did not want to embark on process and organizational restructuring, possibly

because in the meantime much of the work had become sufficiently routinized that it could be

performed with automation.

The development of additive manufacturing (AM) technologies has made the calculation on

how, where, and what to produce more difficult. For example, some have argued that AM results in a

clash between the classes of heuristics that favor simplicity and those that favor complexity. In the case

of traditional manufacturing, which is optimized for modularization, “simple is better” is a powerful

heuristic, which contrasts with the “complexity for free” motto of AM. Thus, AM and traditional

manufacturing may be substitutes in some instances, but they are also complements, as AM is used for

prototyping or production in small batches, whereas traditional mass production (e.g., using advanced,

sophisticated, AI-controlled robots) is used for production in larger volume (Pedota and Piscitello,

2021).

One of the putative benefits of Industry 4.0 was the belief that current offshoring and relocation

strategies would be reconsidered. Although the adoption of Industry 4.0 technologies may have

reduced the pace of delocalization to lower-cost countries, it is still too early to observe a positive

impact of increasing digitalization on reshoring (Ancarani and Di Mauro, 2018).4

More recently, platforms for contracting gig workers have emerged and perform the role of

intermediary between the contractor and contractee. These platforms were expected to enable

outsourcing and offshoring of a variety of tasks (e.g., Hong and Pavlou, 2017). Although these gig

work platforms initially grew rapidly, their growth seems to have slowed, if one can generalize from

4
Given the difficulties that many countries faced with global supply chains during the 2020-21 COVID-19 outbreak, some

are predicting that this will trigger reshoring (Barbieri et al., 2020).
17
the revenue of the largest among them, Upwork. Like offshoring, platform-organized gig work is a

digitally enabled method of hiring labor. However, it is possible that in the long run much of this

outsourced work will be automated. Although, in principle, online outsourcing platforms provide

ubiquitous access with no geographic constraints, country differences continue to matter even in

software development and IT services (Carmel and Tijia, 2005).

Another fundamental development due to digitalization is the introduction and increasing

economic and social centrality of online digital platforms (Kenney and Zysman 2016; Van Dijck 2013;

Kenney et al., 2020). Online platforms are restructuring business sectors and firm operations for an

increasingly large percentage of the workforce and thus increasing their vulnerability to reorganizations

enabled by adoption of these platforms. We know little about which skills affect the ability to earn an

income through a platform. In the case of Uber, Deliveroo, and other related transportation platforms,

these skills may be minimal. However, a successful Instagram influencer or YouTuber certainly has

developed skills, though they are quite different from previous skills. Platforms such as the Apple

Appstore and Google Play provide resources, such as software development kits, application

programming interfaces, and access to a large number of potential customers for entrepreneurial

innovators, but we know little about the skills that these “platform-dependent entrepreneurs” possess

and require to successfully sustain their precarious income (Cutolo and Kenney, 2021; Cutolo et al.

2021).

These changes caused by digitalization are illustrated by changes in services that are as intrinsically

human as answering a customer query at a call center. Only fifteen years ago, nearly all the activities

related to answering a verbal customer query were handled by people, and cost reduction was achieved

by offshoring to a lower-cost developing country (see, e.g., Dossani and Kenney, 2007). Today, the

initial contact is handled with voice recognition software (i.e., a service “bot”) that locates the

18
customer’s account and offers to provide various types of information or to route the call to the

appropriate department.

To summarize, digital technologies and skills are increasingly intertwined, integrating

complementary physical, intangible, and computational technologies that require multiple and varied

skills, from STEM to interpersonal and social skills.

4. Three Challenges

As digital technologies coevolve with firm innovation and skills, and firm reorganize their

activities, the short review above leads to more questions than answers about their impact on

economies and societies. We summarize these questions in terms of three research challenges:

19
 1. Enhancement of understanding of the main future trajectories of digital technologies

2. Increasing understanding and improving data about the adoption, use, and recombination of

digital technologies

3. Assessment of their impact on productivity and inequality.

Building on Schumpeter (1939) and Kondratiev’s theory of long waves, Chris Freeman stated

that different historical periods (following the first Industrial Revolution) are led by the development

and diffusion of fundamental technologies that shape economies and societies (Freeman, 1982, 1991).

The latest of these revolutions can be dated to the 1970s, when ICT began to be adopted more generally

(Freeman and Louçã, 2002; Freeman and Perez, 1988). Although there is little doubt that ICT has

changed industries, production organization, the process of work, the types and characteristics of

employment, and the requisite skills (Bresnahan and Trajtenberg, 1995), the discussion in Sections 2

and 3 shows that contemporary digital technologies only partly follow the earlier patterns in which

alternative, competing technologies were developed until a dominant design emerged, after which the

pace of change slowed and involved mostly process and incremental innovations.

Although the computer industry has exhibited some life-cycle characteristics (Malerba et al.,

1999), in many respects, it has violated them, as with each new wave of digitalization, more computers

are deployed, and they display remarkable new capabilities. This is largely because of Moore’s law,

which postulated a log-linear relationship between circuit density, an ever-lower cost per transistor, and

relatively predictable, yet astonishingly rapid advances in computing power over time.

20
 This increase in data processing capacity created continual new opportunities for innovation, as

some things now could be represented digitally that never could before. The key unknown in this

technological advance is what the constantly increasing processing power would make feasible, as

software transforms processing power into applications in the real world. For example, as recently as a

decade ago, a self-driving car was improbable because the ability to sense and process all the necessary

variables in the environment was not yet developed. Thus, at the time, the skills a human driver could

display were impervious to replacement as well as automation (Autor et al., 2003). In 2021, some of

individual skills, such as collision detection and lane-change warnings, can be performed by

automobiles, though human monitoring and, at critical moments, human decision-making remain

necessary. The advent of fully autonomous driving represents the automation of the entire activity—

unless and until, of course, those vehicles confront entirely unpredictable environmental conditions that

its AI cannot recognize and handle.

The first challenge that we face is understanding and predicting the trajectories of digital

technologies, due to the complex and nonlinear relation and interdependence among digital

technologies, organizations, and skills (Archibugi, 2017; Arthur, 2009; 2013). Digital technologies are

constantly evolving and do not seem to have settled into a well-defined trajectory yet. Therefore, we

seem to be in the midst of an exploratory phase of a new technology, which makes it difficult to predict

the next emerging technology or even how it will be defined (Chiarello et al., 2018). In their first AI

index report, Shoham et al. (2017, p7) stated: “Without the relevant data for reasoning about the state

of AI technology, we are essentially ‘flying blind’ in our conversations and decision-making related to

AI.” Despite temporary signs of convergence and “cooling down” of the AI trajectory (Klinger et al.,

2020), the wide range of potential applications, current low rates of adoption, and numerous

opportunities for entrepreneurs to enter the market with new business models (Autio et al., 2018) may

require us to revisit our understanding and models of past technological revolutions, in order to
21
understand the future development of digital technologies. The various digital technologies are

evolving at different rates and following different trajectories. Although it is too early to call them

revolutionary, some of these technologies have the potential to become so. However, at present they

can be differentiated, as some show the characteristics of general purpose technologies (e.g., cloud

computing), whereas others appear to have a limited scope and thus can be understood as enabling

technologies (e.g., 3D printing) (Martinelli et al., 2021).

22
 The second and related challenge concerns the fact that, although we have seen massive

adoption of earlier digital technologies, such as computers and computer-controlled machines, the

extent to which more recent waves of digital technologies and automation are being adopted is unclear.

For instance, the Digital Technology Module of the 2008 United States Census Bureau Annual

Business Survey (table 3a) shows that, in the US, technologies such as augmented reality, autonomous

vehicles, machine learning, and machine vision software have been adopted by between 0.9 and 4.5

percent of the firms, though this includes those that are merely testing them (see also Vannuccini and

Prytkova, 2020). In Europe, Deng, Plümpe, and Stegmaier (2020) ascertain from data in the IAB

Establishment Survey that in 2018 only 1.55% of German companies had adopted robots. Similarly, the

AI index 2019 Annual Report (Perrault et al., 2019) shows that the industry with the largest share of

AI-related jobs, information, had only 2.4% of the AI-related vacancies. Of course, these are only

estimates, and firms may not even be aware that they are using the outputs of AI analysis. Consider, for

example, a firm that is advertising on Google Search or YouTube. The target of its advertisement may

have been chosen with a machine learning algorithm, thus unwittingly they are users. By paying for the

advertisement, is the firm using AI? As with all intangibles, if AI is embedded in software, but is not a

“physical thing,” then it is difficult to measure its adoption, even though it may be substantial (Marcus

and Davis, 2019). A fundamental difficulty in measuring the adoption and diffusion of digital

technologies is that they have become increasingly intangible or embody elements that are intangible

and therefore require data about the value of the transactions involved in accessing a service or the

technologies (e.g., a software), which is complicated to capture.

If, as it seems, adoption of new digital technologies is low, why is that the case? Are firms

waiting to see how digital technologies develop? Are there too many different technologies that have

not yet fully evolved or, perhaps, are changing too rapidly? Is the technology moving too rapidly for

firms and workers to adapt to new routines and skills?

(including on occupations, as the variations in predictions suggest), unless we improve our ability to

study their evolution and adoption. At the aggregate level, since the digital technological revolution,

high-income economies have experienced a secular slowdown in productivity growth (Baldwin and

Teulings, 2014; Gordon, 2015). Although this may be linked to measurement issues in the prevalent

service industries, the reduced gross domestic product per capita growth rates suggest that, unlike in

earlier technological revolutions, well-being, measured in economic terms, is not growing. Stimulated

by the ICT productivity paradox in the 1980s, the adoption of new technologies is not sufficient for

increased productivity—it has to be accompanied by transformations in the firm, jobs, and skills

(Brynjolfsson et al., 2017).

In the few decades since the digital revolution began, productivity growth has continued to

decline, on average. It has been argued that, although average productivity growth has declined

throughout the digital age, except during the financial crisis, the top performers have experienced

sustained productivity growth (Andrews et al., 2016; Haldane, 2017). This again raises the possibility

that only a few firms can successfully integrate digital technologies, whereas the average firm cannot

do so effectively. Presenting an alternative explanation, Brynjolfsson and Collis (2019) argue that

digital technologies may be having a negative effect on productivity growth because they offer so many

“free” goods and services. As noted in the example above, the productivity created by Google Search,

Maps, and YouTube is captured only in the advertising served to users, as the services are free.

Similarly, smartphones now have high-quality cameras, watches, compasses, fitness monitors, and

many other products/services at a comparatively low price. Moreover, new apps appear constantly and

can be downloaded at a very low or even no cost. Thus, even though the true value of the goods and

services the consumer received increased, productivity may have dropped.

24
 The digital age has also been accompanied by a steep increase in inequality (Atkinson, 2015;

Piketty, 2014), which has been explained in many ways. For example, labour compensation

(Karabarbounis and Neiman, 2013) and unionization (Freeman, 2007) have declined, but the

compensation of top earners has dramatically increased (Atkinson, Piketty, and Saez, 2011), as has firm

size (Mueller, Ouimet, and Simintzi, 2017) and industrial concentration (Autor et al., 2020). Some

researchers have found that differences between firms explain two-thirds of the earning variability

across workers (Card et al., 2018; Song et al., 2019), suggesting that this outcome is based, in part, on

interfirm competition. How much of this is directly related to digitalization is difficult to measure,

although researchers have found a higher concentration in patenting (Forman and Goldfarb 2020) and

that AI adoption may have increased the concentration of intangible assets in fewer firms (Rock 2019;

Tambe et al. 2019). Inequality is also observed geographically. For example, Bloom et al. (2020) study

the geographical development of new technologies, primarily but not exclusively digital, and their

accompanying employee skill bases. They found that, as employment in these new technologies grew,

it was initially concentrated in local hubs, before spreading geographically. Yet the initial hubs retained

a disproportionate share of employment, particularly in the most desirable highly skilled positions.

If digital technologies keep evolving and do not settle into well-defined paradigms (Dosi,

1982), definitively measuring their recombination, adoption, and use is likely to remain difficult. We

need substantially more evidence, at the firm, sector, and country level, to better understand how digital

technologies are evolving and how firms, workers, and their routines and skills coevolve with them.

This conclusion aligns with the call voiced by policy makers, academics, and other stakeholders for

more and better data that can create insights into the implications of these technologies for economies

and societies (National Academies of Sciences Engineering and Medicine, 2017; OECD, 2017; Raj and

Seamans, 2018).

25
 The papers in this special issue address some of these challenges directly.

5. The Papers in the Special Issue

26
 The papers in this special issue offer significant insight into the challenges identified above but

also suggest that there is ample room for more research as we seek to understand the rapidity with

which digital technologies are evolving and how they are transforming work.

The first two papers discuss the second and third challenges, respectively, as they explore new

ways to measure firm investment in capital that increases digitization (Harrigan, et al., in this issue) and

automation (Domini et al., in this issue), and study their impact on employment. Harrigan et al. (in this

issue) explore how an increasing share of workers with STEM skills and experiences (“techies”),

associated with an increase in digitalization, changes the composition of employment at a firm. They

show that techies contribute to driving technological change within firms, resulting in greater job

polarization within industries, rather than within the firm. This is because of the recomposition within

industries (firms with more techies and digital technologies increase their market share).

Domini et al. (in this issue) employ firm transaction data to measure changes in automation due

to investment in automation-intensive capital goods. They then study the impact of these automation

“spikes” on employment. In contrast to Harrigan et al., they find that capital investment does not seem

to change the composition of labor (i.e., the techies are not complementary to capital), but, in line with

Harrigan et al., they see that, after automation spikes, overall employment at automating firms

increases. So, consistent with Harrigan et al., they claim that automation helps increase the focal firm’s

market share (Acemoglu et al., 2020), and the impact on employment is negative at the industry level.

27
 As discussed earlier, technology, organizations, and skills co-evolve. The paper by Cirillo et al.

(in this issue) focus on the specific relationship between technology and skills. Employing an original

dataset on occupations in Italy in the period 2011-2016, they analyze the relationship between

digitalization, task routinization, and employment. They find that “digital occupations” (i.e., activities

that explicitly involve digital technologies) are not necessarily associated with routinized occupations

and that, although employment in digital occupations is growing, jobs in routinized occupations tend to

decrease.

Discussing the challenge of attaining a better understanding of the relationship between the

evolution of digital technologies, organization, and skills, Goos, Rademakers, and Röttger (in this

issue) analyze data on workers collected before and after a plant closure to examine what happens to

redundant workers. They show that workers who engaged in routine-intensive tasks, in contrast to their

peers who performed less-routinized tasks, struggled to find employment at new factories. And workers

who had less-routinized jobs but worked at older factories could not find jobs of similar quality,

suggesting that their learning at an older factory did not prepare them for working with changing

(digital) technologies in new factories.

28
 The question of which skills workers should develop to keep up with increasing digitalization is

somewhat vexatious. Two papers in the special issue present significant evidence that is relevant to this

question, addressing the challenge of achieving a better understanding of the evolution of digital

technologies, organization, and skills, from the perspective of skills. In a cross-country data setting,

Falck et al. (in this issue) show that workers with higher ICT skills earn more than their peers with

lower ICT skills and gain experience working with more abstract content (i.e., less routinized and using

more advanced digital technologies). Further light on these issues is cast by Black et al who use data

from a survey of US high school graduates in 1982, at the inception of the digital revolution. They

found that the educational experience that was most important in preparing them for a labor market in

which they would need to use and learn new digital technologies, was taking courses in advanced

mathematics. Those who did so were more likely to find better-paid jobs, better able to adapt to

technological evolution, and more likely to be employed in STEM occupations.

The next paper in the issue addresses the third challenge, the role of digital technologies in the

economy and society. Neil Foster-McGregor and Bart Verspagen (in this issue) construct different

trade scenarios to study the role of trade relations between countries (including offshoring) in the risk

of job replacement by automation in any of those countries. In contrast to the micro-level evidence

described by Harrigan et al. (in this issue), they find that a country’s sectoral structure largely explains,

the risk that its jobs may be replaced by automation, i.e., automation risk varies more between sectors

than within them, across countries, and this depends on the composition of the country’s industrial

base. They find a negative relationship between the risk of automation and labor productivity. Further,

they find that trade increases the automation risk for high-productivity European countries, which is

caused by trade between European and non-European countries. These European countries do not

offshore automation risk but, instead, import it. This risk is concentrated in manufacturing, trade,

29
transport, and finance. In other words, in high-productivity countries, trade (and offshoring) seems to

shift the structure of employment toward activities that are more likely to be automated.

6. Conclusions

The large discrepancies in the literature in estimating the potential impact of automation and

some digital technologies on employment suggest that predicting the skills needed for future

employment is filled with uncertainty. Digitalization is often more about the displacement and deep

transformation of activities and their organization than about a simple one-to-one replacement of jobs.

These transformations are often unforeseen and coevolve with innovation routines and skills. For

instance, whereas existing skills are threatened by innovation, and the repetitive nature of tasks

increased somewhat, as did the use of software, an examination of the changes in the O*NET

occupational skills from 2005 to 2015 found only modest changes in occupational skills (Freeman et

al., 2020).

One of the difficulties in predicting the impact of digital technologies on skills is their plasticity

and the ease of introducing and diffusing new software. This means that a broad and continuous wave

of constant experimentation is underway as software becomes the core of nearly every activity. In

Garud et al.’s (2008) reading, software is “incomplete by design” and thus open to constant revision.

This can be seen in many industries. For example, farming is being transformed by digital

technologies, so simply training farmers to program farm machinery will not halt the displacement of

farmers or farm workers. It is more likely that the farmer (owner) will become a higher-level decision-

maker about when to upgrade software and equipment, a decision that requires an understanding of

accounting as much as software skills (Kenney et al., 2020). The other impact of automation may be

that it will allow farmers to reduce the number of employees hired by replacing them with a
30
combination of hardware and software. The most difficult task for farmers, and thus a skill that they

will have to learn, is determining when the software is making bad decisions or recommendations. 5 The

difficulty of responding to some types of technological change in very skilled professions can be

illustrated by the case of airline pilots who are unlikely to be able to ward off competition from drones

through obtaining more advanced training as pilots or learning to write software. It is easy to agree that

everyone should be a lifelong learner, but the question is: what should everyone be learning, and what

is the likelihood that doing so will lead to employment enabling a middle-class income?

The already turbulent and unpredictable evolution of digitalization was made even more

uncertain by the COVID-19 pandemic. As the preparation of this Special Issue proceeded, almost

overnight the pandemic changed the physical location of many types of jobs, as those who could work

remotely, from home, were told to do so. This radically disrupted the physical organization of work, as

meetings and other in-person service activities were performed online. The pandemic accelerated the

need for digitalization to enable organizations to continue functioning in the short term as well as

increase resilience in the long term. It created a further impetus to study the impact of digitalization on

the supply of and demand for skills, the mismatch in relevant skills, shortages of skilled workers, and

related needs (Fink, 2020).

The pandemic has had a profound impact on all types of work. For example, performing some

kinds of work offsite was made possible with digital technologies such as Slack, Zoom, and Ding Talk

(China). Like any mass adoption of a new communication medium, this one certainly changed some

5
The automation literature has little consideration of how the software may be programmed to favor particular actors and

thus express the desires of its programmers. Consider, for example, failure-prediction sensors on a piece of machinery.

These can be vitally important in preventing dangerous or costly equipment breakdowns. Of course, they could also be

programmed to predict failure earlier than necessary, thereby increasing replacement costs and the profits of the equipment

maker.
31
employee skill sets, career paths, and created entirely new innovation paths. It is uncertain which of

them will survive/expand in the post-pandemic period, but business exploration in many organizations

was initiated during the pandemic. In addition to affecting services, the pandemic altered physical

production and distribution, as factories and warehouses became an important locus of disease

transmission, creating increased pressure for automation (Aratani, 2020; Bunge and Newman, 2020).

This special issue was produced to encourage more research that would contribute to a better

understanding of how digitalization affects innovation and skills, by improving our understanding of

the coevolution between digital transformation, innovation routines and processes, and changes in the

skills required to innovate, learn, and adapt—and how these relationships help transform the

organization of firms and industries. The papers in this issue present relevant findings on understanding

these three dyadic relationships. To address the challenges laid out in Section 5, we call for future

research that incorporates the endogenous features of these three dynamics (digital technologies,

innovation, and skills). This is probably best achieved by combining multiple or cross levels of

analysis, embracing ideas and concepts in multiple disciplines, to define a new set of stylized facts on

how innovation practices are being transformed, with their implications for work, jobs, and skills.

32
REFERENCES

Acemoglu, D., D. Autor, J. Hazell, Restrepo, P., 2020. AI and jobs: Evidence from online vacancies.

NBER Working Paper Series, No. 28257 (downloaded from

https://economics.mit.edu/files/21056/).

Acemoglu, D., Restrepo, P., 2018. The race between man and machine: Implications of technology for

growth, factor shares, and employment. American Economic Review 108(6), 1488-1542.

Acemoglu, D., Restrepo, P., 2019. The wrong kind of AI? Artificial intelligence and the future of labor

demand. NBER. Cambridge, MA. https://doi.org/10.3386/w25682.

Adermon, A., Gustavsson, M., 2015. Job polarization and task‐biased technological change: Evidence

from Sweden, 1975–2005. Scandinavian Journal of Economics, 117(3), 878-917.

Agrawal, A., Gans, J., Goldfarb, A., 2018. Prediction Machines: The Simple Economics of Artificial

Intelligence. Harvard Business Press, Boston.

Agrawal, A., Gans, J., Goldfarb, A., 2019. Artificial intelligence: The ambiguous labor market impact

of automating prediction. Journal of Economic Perspectives 33(2), 31–50.

Alcácer, J., Cantwell, J., Piscitello, L., 2016. Internationalization in the information age: A new era for

places, firms, and international business networks? Journal of International Business Studies

47(5): 499-512.

Andrews, D., Criscuolo, C., Gal, P. N., 2016. The best versus the rest: The global productivity

slowdown, divergence across firms and the role of public policy. OECD: Paris (downloaded from

https://www.oecd-ilibrary.org/content/paper/63629cc9-en).

33
Aratani, L., 2020. Robots on the rise as Americans experience record job losses amid pandemic.

Guardian (November 27) https://www.theguardian.com/technology/2020/nov/27/robots-

replacing-jobs-automation-unemployment-us

Ancarani, A., Di Mauro, C., 2018. Reshoring and industry 4.0: How often do they go together? IEEE

Engineering Management Review 46(2), 87-96.

Archibugi, D., 2017. Blade runner economics: Will innovation lead the economic recovery? Research

Policy 46(3),535–43.

Arntz, M., Gregory, T., Zierahn, U., 2017. Revisiting the risk of automation. Economics Letters 159,

157-160.

Arthur, W. B., 2013. Complexity economics: A different framework for economic thought. Paper # 13-

04-012. Santa Fe Institute (downloaded from https://www.santafe.edu/research/results/working-

papers/complexity-economics-a-different-framework-for-eco/).

Arthur, W.B., 2009. The Nature of Technology: What It Is and How It Evolves. Simon and Schuster,

New York.

Atkinson, A., 2015. Inequality: What Can Be Done? Harvard University Press, Cambridge, MA.

Atkinson, B., Piketty, T., Saez, E., 2011. Top incomes in the long run of history. Journal of Economic

Literature 49(1), 3–71.

Autio, E., Nambisan, S., Thomas, L. D., Wright, M., 2018. Digital affordances, spatial affordances, and

the genesis of entrepreneurial ecosystems. Strategic Entrepreneurship Journal, 12(1), 72-95.

34
Autor, D., 2014. Polanyi's paradox and the shape of employment growth. NBER Working Paper Series,

No. 20485 (downloaded from https://www.nber.org/papers/w20485/).

Autor, D., 2015. Why are there still so many jobs? The history and future of workplace automation.

Journal of Economic Perspectives 29(3), 3-30.

Autor, D., Dorn, D., Hanson, G., 2015. Untangling trade and technology: Evidence from local labour

markets. Economic Journal, 125(584), 621-646.

Autor, D., Dorn, D., Katz, L.F., Patterson, C., Van Reenen, J., 2020. The fall of the labor share and the

rise of superstar firms. Quarterly Journal of Economics 135(2), 645–709.

Autor, D., Katz, L. F., Krueger, A., 1998. Computing inequality: Have computers changed the labor

market? Quarterly Journal of Economics, 113(4), 1169-1213.

Autor, D., Levy, F., Murnane, R., 2002. Upstairs, downstairs: computers and skills on two floors of a

large bank. ILR Review 55(3), 432-447.

Autor, D., Levy, F., Murnane, R., 2003. The skill content of recent technological change: An empirical

exploration. Quarterly Journal of Economics, 118(4), 1279-1333.

Autor, D., Salomons, A., 2018. Is automation labor-displacing? Productivity growth, employment, and

the labor share. Brookings Papers on Economic Activity Conference Drafts, March 8-9, 2018, 1-

72.

35
Baldwin, R., Teulings, C., 2014. Secular stagnation: Facts, causes and cures. Center for Economic

Policy Research (downloaded from https://voxeu.org/content/secular-stagnation-facts-causes-

and-cures/).

Barbieri, P., Boffelli, A., Elia, S., Fratocchi, L., Kalchschmidt, M., Samson, D., 2020. What can we

learn about reshoring after Covid-19? Operations Management Research 13(3), 131-136.

Barley, S.R. and Kunda, G., 2011. Gurus, Hired Guns, and Warm Bodies. Princeton: Princeton

University Press.

Barth, E., Davis, J.C., Freeman, R.B., McElheran, K., 2020a. Twisting the Demand Curve:

Digitalization and the Older Workforce. NBER Working Paper, No. 28094 (downloaded from

https://ide.mit.edu/sites/default/files/publications/w28094.pdf).

Barth, E., Roed, M., Schøne, P., Umblijs, J., 2020b. How robots change within-firm wage inequality.

IZA Institute of Labor Economics, No. 13605. (downloaded from

https://www.iza.org/publications/dp/13605/how-robots-change-within-firm-wage-inequality/).

Bessen, J., 2019. Automation and jobs: When technology boosts employment. Economic Policy

34(100), 589-626.

Black, S. E., Muller, C., Spitz-Oener, A., He, Z. Hung, K. and Warren, J. R. 2021. The importance of

STEM: High school knowledge, skills and occupations in an era of growing inequality. Research

Policy in this issue.

36
Bloom, N., Hassan, T.A., Kalyani, A., Lerner, J., Tahoun, A., 2020. The geography of new

technologies. Institute for New Economic Thinking, Working Paper No. 126 (downloaded from

https://www.ineteconomics.org/uploads/papers/WP_126-Tahoun-et-al.pdf).

Bresnahan, T., Brynjolfsson, E., Hitt, L., 2002. Information technology, workplace organization, and

the demand for skilled labor: Firm-level evidence. Quarterly Journal of Economics, 117(1), 339-

376.

Bresnahan, T., Trajtenberg, M., 1995. General purpose technologies “engines of growth.” Journal of

Econometrics 65 (1): 83–108.

Brynjolfsson, E., Collis, A., 2019. How should we measure the digital economy? Harvard Business

Review 97(6), 140-148.

Brynjolfsson, E., McAfee, A., 2014. The Second Machine Age: Work, Progress, and Prosperity in a

Time of Brilliant Technologies. W.W. Norton, New York.

Brynjolfsson. E., Mitchell, T., 2017. What can machine learning do? Work-force implications. Science

358(6730), 1530-1534.

Brynjolfsson, E., Rock, D., and Syverson, C., 2017. Artificial intelligence and the modern productivity

paradox: A clash of expectations and statistics. NBER Working Paper No. 24001. Cambridge,

MA. https://doi.org/10.3386/w24001.

Bunge, J., Newman, J., 2020. Tyson turns to robot butchers, spurred by coronavirus outbreaks. Wall

Street Journal, (July 9) (downloaded from https://www.wsj.com/articles/meatpackers-covid-

safety-automation-robots-coronavirus-11594303535/).

37
Card, D., Cardoso, A.R., Heining, J., Kline, P., 2018. Firms and labor market inequality: Evidence and

some theory. Journal of Labor Economics 36(S1), 13–70.

Carmel, E., Tjia, P. 2005. Offshoring information technology: Sourcing and outsourcing to a global

workforce. Cambridge University Press, Cambridge.

Cedefop, 2018. Insights into skill shortages and skill mismatch: Learning from Cedefop’s European

skills and jobs survey. Cedefop. DOI 10.2801/897740

Chiarello, F., Trivelli, L., Bonaccorsi, A., Fantoni, G., 2018. Extracting and mapping industry 4.0

technologies using Wikipedia. Computers in Industry 100, 244–57.

Cirillo, V., Evangelista, R., Guarascio, D., Sostero, M., 2020. Digitalization, routineness and

employment: An exploration on Italian task-based data.” Research Policy, No. 104079

(downloaded from https://www.sciencedirect.com/science/article/abs/pii/S0048733320301578).

Cutolo, D., Hargadon, A.B., and Kenney, M. 2021. Competing on platforms: Recognizing and

navigating the risks. Sloan Management Review (March), 1-8.

Cutolo, D., Kenney, M., 2021. Platform-dependent entrepreneurs: Power asymmetries, risks, and

strategies in the platform economy. Academy of Management Perspectives. In Press.

Colfer, L.J., Baldwin, C.Y., 2016. The mirroring hypothesis: Theory, evidence, and exceptions.

Industrial and Corporate Change 25, 709–738.

Das, S., Steffen, S., Clarke, W., Reddy, P., Brynjolfsson, E., Fleming, M., 2020. Learning occupational

task-shares dynamics for the future of work. In AIES 2020: Proceedings of the AAAI/ACM

Conference on AI, Ethics, and Society, 36–42. New York: ACM.

https://doi.org/10.1145/3375627.3375826

38
Deming, D., 2017. The growing importance of social skills in the labor market. Quarterly Journal of

Economics 132(4), 1593-1640.

Deng, L., Plümpe, V., Stegmaier, J., 2020. Robot adoption at German plants. Institut für

Wirtschaftsforschung Halle, No. 19 (downloaded from

https://www.iwh-halle.de/publikationen/detail/robot-adoption-at-german-plants/).

De Stefano, V., 2015. The rise of the just-in-time workforce: On-demand work, crowdwork, and labor

protection in the gig-economy. Comparative Labor Law and Policy Journal 37, 471-504.

Domini, G., Grazzi, M., Moschella, D., Treibich, T., 2020. Threats and opportunities in the digital era:

Automation spikes and employment dynamics. Research Policy, No. 104137 (downloaded from

https://www.sciencedirect.com/science/article/abs/pii/S0048733320302122?dgcid=rss_sd_all).

Dosi, G., 1982. Technological paradigms and technological trajectories: A suggested interpretation of

the determinants and directions of technical change. Research Policy 11(3), 147–62.

Dossani, R., Kenney, M., 2007. The next wave of globalization: Relocating service provision to

India. World Development 35(5), 772-791.

Esteva, A., Kuprel, B., Novoa, R.A., Ko, J., Swetter, S.M., Blau, H.M., Thrun, S., 2017.

Dermatologist-level classification of skin cancer with deep neural networks. Nature 542(7639),

115-118.

Eurofound, 2018. Automation, digitisation and platforms: Implications for work and employment,

Publications Office of the European Union, Luxembourg (downloaded from

https://www.eurofound.europa.eu/sites/default/files/ef_publication/field_ef_document/

ef18002en.pdf).

39
Eurofound, 2020. Labour market change: Trends and Policy approaches towards flexibilisation.

Eurofound (downloaded from

https://www.eurofound.europa.eu/publications/flagship-report/2020/labour-market-change-

trends-and-policy-approaches-towards-flexibilisation/).

Falck, O., Heimisch-Roecker, A., and Wiederhold, S., 2021. Returns to ICT skills. Research Policy in

this issue

Felten, E., Raj, M., Seamans, R., 2018. A method to link advances in artificial intelligence to

occupational abilities. AEA Papers and Proceedings 108, 54–57.

Felten, E., Raj, M., Seamans, R., 2019. The effect of artificial intelligence on human labor: An ability-

based approach. Academy of Management. DOI: 10.5465/AMBPP.2019.140

Fink, L., 2020. Conducting information systems research in the midst of the covid-19 pandemic:

Opportunities and challenges. Information Systems Management 37, 256-259.

Forman, C., Goldfarb, A., 2020. Concentration and agglomeration of IT innovation and

entrepreneurship: Evidence from patenting. National Bureau of Economic Research, Cambridge,

MA. https://doi.org/10.3386/w27338.

Foster-McGregor, N. and Verspagen, B. 2021. Job automation risk, economic structure and trade: A

European perspective. Research Policy. In this issue

Frank, M., Autor, D., Bessen, J., Brynjolfsson, E., Cebrian, M., Deming, D., Feldman, M., Groh, M.,

Lobo, J., Moro, E., Wang, D., Youn, H., Rahwan, I., 2019. Toward understanding the impact of

artificial intelligence on labor. Proceedings of the National Academy of Sciences of the United

States of America 116(14), 6531–6539.

Freeman, C., 1982. Innovation and long cycles of economic development. International Seminar on
40
 Innovation and Development at the Industrial Sector (downloaded from

https://enterrasolutions.com/media/docs/2013/02/JoseCassiolato_2.pdf).

Freeman, C., 1991. Innovation, changes of techno-economic paradigm and biological analogies in

economics. Revue Economique 42(2), 211-231.

Freeman, C., Louçã, F., 2002. As Time Goes by: From the Industrial Revolutions to the Information

Revolution. Oxford University Press, Oxford.

Freeman, C., Perez, C., 1988. Structural crises of adjustment: business cycles and investment behavior,

in: Dosi, G., Freeman, C., Nelson, R.R., Silverberg, G., Soete, L.L.G. (Eds.), Technical Change

and Economic Theory, pp. 38-66. Pinter, London.

Freeman, R., 2007. Labor market institutions around the world. 13242. NBER Working Paper No.

13242. Cambridge, MA. https://doi.org/10.3386/w13242.

Freeman, R., Ganguli, I., Handel, M., 2020. Within-occupation changes dominate changes in what

workers do: A shift-share decomposition, 2005–2015. AEA Papers and Proceedings 110, 394-99

(downloaded from https://www.aeaweb.org/articles?id=10.1257/pandp.20201005/).

Frey, C., Osborne, M., 2017. The future of employment: How susceptible are jobs to computerisation?

Technological Forecasting and Social Change 114, 254-280.

Garud, R., Jain, S., Tuertscher, P., 2008. Incomplete by design and designing for incompleteness.

Organization Studies, 29(3), 351-371.

Goos, M., Manning, A., 2007. Lousy and lovely jobs: The rising polarization of work in Britain.

Review of Economics and Statistics, 89(1), 118-133.

41
Goos, M., Manning, A., Salomons, A., 2009. Job polarization in Europe. American Economic

Review, 99(2), 58-63.

Goos, M., Manning, A., Salomons, A., 2014. Explaining job polarization: Routine-biased technological

change and offshoring. American Economic Review 104(8), 2509-26.

Goos, M., Rademakers, E., Röttger, R., 2020. Routine-biased technical change: Individual-level

evidence from a plant closure. Research Policy, No. 104002 (downloaded from

https://www.sciencedirect.com/science/article/abs/pii/S0048733320300822).

Gordon, R., 2015. Secular stagnation: A supply-side view. American Economic Review 105(5), 54-59.

Haldane, A., 2017. Productivity puzzles. Speech, London School of Economics, March 17.

Harrigan, J., Reshef, A., and Toubal, F., 2020. The march of the techies: Job polarization within and

between firms. Research Policy, No. 104008 (downloaded from

https://www.sciencedirect.com/science/article/abs/pii/S0048733320300871).

Helper, S., Mertens, R., Seamans, R., 2019. Who profits from industry 4.0? Theory and evidence from

the automotive industry. (January 31). NYU Stern School of Business (downloaded from

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3377771/).

Henfridsson, O., Nandhakumar, J., Scarbrough, H., Panourgias, N., 2018. Recombination in the open-

ended value landscape of digital innovation. Information and Organization 28(2), 89-100.

Hitt, L., 1999. Information technology and firm boundaries: Evidence from panel data. Information

Systems Research 10(2), 134-149.

42
Hinings, B., Gegenhuber, T., Greenwood, R., 2018. Digital innovation and transformation: An

institutional perspective. Information and Organization 28(1), 52-61.

Hong, Y., Pavlou, P.A., 2017. On buyer selection of service providers in online outsourcing platforms

for IT services. Information Systems Research 28(3), 547–562.

Kagermann, H., 2015. Change through digitalization: Value creation in the age of Industry 4.0, in:

Albach, H., Meffert, H., Pinkwart, A., Reichwald, R. (Eds.), Management of Permanent Change,

pp. 23-45. Springer, Wiesbaden.

Karabarbounis, L., Neiman, B., 2013. The global decline of the labor share. Quarterly Journal of

Economics 129(1), 61–103.

Kauffman, S.A., Levin, S., 1987. Towards a general theory of adaptive walks on rugged landscapes.

Journal of Theoretical Biology 128(1),11–45.

Kenney, M., Massini, S., Murtha, T.P., 2009. Offshoring administrative and technical work: New fields

for understanding the global enterprise. Journal of International Business Studies 40(6), 887-900.

Kenney, M., Serhan, H., Trystram, G., 2020. Digitalization and platforms in agriculture: Organizations,

power asymmetry, and collective action solutions. SSRN (downloaded from

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3638547).

Kenney, M., Zysman, J. 2016. The rise of the platform economy. Issues in Science and

Technology, 32(3), 74-82.

Klinger, J., Mateos-Garcia, J., Stathoulopoulos, K., 2020. A narrowing of AI research? Nesta

(downloaded from https://nesta-test.s3.eu-west-2.amazonaws.com/20201111_narrowing_ai.pdf).

Korinek, A., Stiglitz, J., 2017. Artificial intelligence and its implications for income distribution and

43
 unemployment. NBER Working Paper Series 24174.

Lane, M., Saint-Martin, A., 2021. The impact of artificial intelligence on the labour market: What do

we know so far? OECD Social, Employment and Migration Working Papers, No. 256, Paris,

France.

Lanzolla, G., Pesce, D., Tucci, C.L., 2021. The digital transformation of search and recombination in

the innovation function: Tensions and an integrative framework. Journal of Product Innovation

Management 38(1), 90-113.

Lasi, H., Fettke, P., Kemper, H.G., Feld, T., Hoffmann, M., 2014. Industry 4.0. Business & Information

Systems Engineering 6(4), 239-242.

Lusch, R.F., Nambisan, S., 2015. Service innovation. MIS Quarterly 39(1), 155-176.

Malerba, F., Nelson, R., Orsenigo, L., Winter, S.G., 1999. History-friendly’ models of industry

evolution: The computer industry. Industrial and Corporate Change 8(1), 3–40.

Mann, D.M., Chen, J., Chunara, R., Testa, P.A., Nov, O., 2020. COVID-19 transforms health care

through telemedicine: Evidence from the field. Journal of the American Medical Informatics

Association 27(7), 1132-1135.

Marcus, G., Davis, E., 2019. Rebooting AI: Building Artificial Intelligence We Can Trust. Pantheon,

New York.

Martinelli, A., Mina, A., Moggi, M., 2021 (Forthcoming). The enabling technologies of industry 4.0:

Examining the seeds of the fourth industrial revolution. Industrial and Corporate Change.

Montobbio, F., Staccioli, J., Virgillito M.E., Vivarelli, M., 2020. Robots and the origin of their labour-

saving impact. IZA Discussion Papers, No. 12967 (downloaded from

44
 https://www.iza.org/publications/dp/12967/robots-and-the-origin-of-their-labour-saving-impact).

Mueller, H., Ouimet, P., Simintzi, E., 2017. Wage inequality and firm growth. American Economic

Review, 107(5), 379–83.

Nambisan, S., Lyytinen, K., Majchrzak, A., Song, M., 2017. Digital innovation management:

Reinventing innovation management research in a digital world. MIS Quarterly, 41(1), 223-238.

Nambisan, S., Wright, M., Feldman, M., 2019. The digital transformation of innovation and

entrepreneurship: Progress, challenges and key themes. Research Policy 48(8), 1-9.

National Academies of Sciences Engineering and Medicine, 2017. Information technology and the U.S.

workforce: Where are we and where do we go from here? National Academies Press,

Washington, DC.

Nedelkoska, L., Quintini, G., 2018. Automation, Skills Use and Training. OECD, Paris.

Pedota, M.J., Piscitello, L., 2020. Additive manufacturing and creativity enhancement: An undervalued

form of complementarity, 21st CINet Conference, 20-22 September 2020, Milan.

O’Donovan, C., Smith, A., 2020. Technology and human capabilities in UK makerspaces. Journal of

Human Development and Capabilities 21(1), 63–83.

OECD, 2017. The next production revolution: Implications for governments and business. OECD

(downloaded from https://www.oecd.org/governance/the-next-production-revolution-

9789264271036-en.htm).

Perrault, R., Shoham, Y.Y., Brynjolfsson, E., Clark, J., Etchemendy, J., Grosz, B., Lyons, T., Manyika,

J., Mishra, S., Niebles, J.C., 2019. The AI Index 2019 Annual Report. Human-Centered Artificial

Intelligence Institute. Stanford University, Stanford, CA.

45
Piketty, T., 2014. Capital in the Twenty-First Century. Belknap Press, Cambridge, MA.

Pouliakas, K., Branka, J., 2020. EU jobs at highest risk of covid-19 social distancing: Is the pandemic

exacerbating the labour market divide? IZA Institute of Labor Economics, Discussion Paper No.

13281 (downloaded from https://www.econstor.eu/bitstream/10419/223723/1/dp13281.pdf).

Raj, M., Seamans, R., 2018. Artificial intelligence, labor, productivity, and the need for firm-level data,

in: Agrawal, A., Gans, J., Goldfarb, A. (Eds.), The Economics of Artificial Intelligence, pp. 553-

65. University of Chicago Press, Chicago.

Rindfleisch, A., O'Hern, M., Sachdev, V., 2017. The digital revolution, 3D printing, and innovation as

data. Journal of Product Innovation Management, 34(5), 681-690.

Rock, D., 2019. Engineering value: The returns to technological talent and investments in artificial

intelligence. SSRN Electronic Journal. https://doi.org/10.2139/ssrn.3427412.

Schumpeter, J.A., 1939. Business Cycles: A Theoretical, Historical and Statistical Analysis of the

Capitalist Process. McGraw Hill, New York.

Schwab, K., 2017. The fourth industrial revolution. World Economic Forum, Geneva, Switzerland.

Shoham, Y., Perrault, R., Brynjolfsson, E., Clark, J., 2017. Artificial Intelligence Index 2017 Annual

Report. Human-Centered Artificial Intelligence Institute. Stanford University, Stanford, CA.

Simon, H., 1969. The Sciences of Artificial. MIT Press, Cambridge, MA.

Simon, H., 2002. Near decomposability and the speed of evolution. Industrial and Corporate Change

11(3), 587–99.

46
Soete, L., 2018. Destructive creation. Explaining the productivity paradox in the digital age, in:

Neufeind, M., O'Reilly, J., Ranft, F. (Eds.), Work and Welfare in the Digital Age: Facing the 4th

Industrial Revolution, pp. 29-46. Rowman & Littlefield, Lanham, MD.

Song, J., Price, D.J., Guvenen, F., Bloom, N., von Wachter, T., 2019. Firming up inequality. Quarterly

Journal of Economics 134(1), 1–50.

Susskind, D., 2017. Re-thinking the capabilities of machines in economics. Department of Economics

Discussion Paper Series, No. 825, University of Oxford, Oxford, England.

Susskind, R., Susskind, D., 2016. Technology will replace many doctors, lawyers, and other

professionals. Harvard Business Review (downloaded from https://hbr.org/2016/10/robots-will-

replace-doctors-lawyers-and-other-professionals).

Tambe, P., Hitt, L., Rock, D., Brynjolfsson, E., 2019. IT, AI and the growth of intangible capital.

SSRN Electronic Journal. https://doi.org/10.2139/ssrn.3416289.

Van Dijck, J., 2013. The Culture of Connectivity: A Critical History of Social Media. Oxford: Oxford

University Press.

Vannuccini, S., Prytkova, E., 2020. Artificial intelligence’s new clothes? From general purpose

technology to large technical system, in: Kanza, S., Frey, J.G., Niranjan, M., Hooper, V. (Eds.),

AI3SD Summer Seminar Series 2020. Southampton, UK (viewed from

DOI:10.5258/SOTON/P0051).

Webb, M., 2019. The impact of artificial intelligence on the labor market. Economics of Innovation

eJournal. DOI: 10.2139/ssrn.3482150

Weil, D., 2014. The Fissured Workplace. Cambridge: Harvard University Press.

47
Wilder, B., Horvitz, E., Kamar, E., 2020. Learning to complement humans. arXiv (downloaded from

https://arxiv.org/abs/2005.00582v1).

Yoo, Y., Boland, R.J., Lyytinen, K., Majchrzak, A., 2012. Organizing for innovation in the digitized

world. Organization Science 23(5), 1398-1408.

Zittrain, J., 2008. The Future of the Internet—and How to Stop It. Yale University Press, New Haven,

CT.

Zolas, N., Kroff, Z., Brynjolfsson, E., McElheran, K., Beede, D.N., Buffington, C., Goldschlag, N.,

Foster, L. Dinlersoz, E., 2021. Advanced technologies adoption and use by U.S. firms: Evidence

from the Annual Business Survey. National Bureau of Economic Research, Working Paper No.

28290 (viewed from https://www.nber.org/system/files/working_papers/w28290/w28290.pdf).

Zuboff, S., 1988. In the Age of the Smart Machine. Basic Books, New York.

48