2. The new challenges for the labour market and the demand for digital skills
This chapter sets the scene for the report by discussing the impact that the digital transition is having on labour markets. The chapter discusses previous empirical evidence that looked into the adoption of digital technologies in jobs and how the adoption of new technologies is reshaping labour market and skill demands. The chapter also briefly reviews the available evidence, produced using online job postings, on the impact of COVID-19 on labour market trends in light of the surge of teleworking. Finally, the chapter discusses some of the most prominent studies that have analysed the impact of digitalisation on the risk of automation and, in turn, on labour markets and employability of workers of all skill levels.
New digital technologies, including artificial intelligence, robotics and information and communication technologies are reshaping the way people live, work and learn. It is nowadays clear that digitalisation presents an immense potential to boost productivity and improve the well-being of all individuals around the world but concerns remain as to whether the digital transition and the future of work will be inclusive for all individuals (OECD, 2019[1]).
While technology is constantly evolving, those changes are poised to replace some of the tasks in jobs that are currently carried out by humans and, in turn, freeing time to produce more innovation, eventually leading to further changes and even more radical shifts in the way humans interact with machines in society and labour markets.
The adoption of new digital technologies can enhance the way people learn, where, when and how they work, and spur their engagement in society by extending everyone’s ability to gather, interpret and analyse information and communicate with others around the globe seamlessly.
Much of these new opportunities are nowadays made possible by the exponentially increasing computing power of digital devices which allows the development and implementation of a range of new digital technologies such as 3D printing, the Internet of Things (IoT) or advanced robotics.
Digital infrastructures and the access to devices such as smartphones has grown, from 4% to 40% of the world’s population in 20 years (OECD, 2017[2]) and interconnected digital devices are used today to collect and distribute big-data from (and to) final users which are then used to optimise all steps of production. The penetration of these new technologies in production as well as in the service sector is showing already substantial repercussions on the way jobs and tasks are carried out.
Some of the areas where digital technologies found significant application are in the production of goods and the delivery of services. As for the former, the introduction of digital technologies in production led to the emergence of the so called “Industry 4.0” (I4.0), or the fourth industrial revolution. The term I4.0 refers to the use in industrial production of interconnected digital technologies that allow processes to be more efficient, timely and cost-saving.1 Among the various digital technologies used in the I4.0 there are the recent developments in machine learning and data science (which allow operating increasingly autonomous and intelligent systems with little or no supervision) to physical sensors that collect and process information that is key to operate the Internet of Things (IoT) and that make second-generation industrial robotics possible.
Among the reasons of their success in industrial production is the fact that machine downtime and repair costs can be greatly reduced when intelligent systems predict maintenance needs. Similarly, significant savings can be had if industrial products can be simulated before being made, and if industrial processes can be simulated before being implemented. Data-driven supply chains greatly speed the time to deliver orders and digital technologies can allow production to be set to meet actual rather than projected demand, reducing the need to hold inventories and lowering failure rates for new product launches (OECD, 2017[2]).
These new trends are becoming increasingly widespread and are affecting parts of the labour market that were not traditionally deemed to be “digital” in the past. Cargo-handling vehicles and forklift trucks, for instance, are nowadays increasingly computerised. Many semi-autonomous warehouses are populated by fast and dexterous robots and complex aspects of the work of software engineers can be performed by algorithms (Hoos, 2012[3])
Digital tools can now replace workers in several routine tasks and even complement workers in tasks that require creativity, problem solving and cognitive skills. The ability of digital technologies to perform some routine tasks led to the vast increase in their use within the service sector. The collection and use of personal information and geolocalised data is used by firms to advertise or tailor their own products in areas like the e-commerce, banking and health.
Just as a matter of example, IBM’s Watson computer can act as a customer service agent (Rotman, 2013[4]) while the Quill programme writes business and analytic reports and Automated Insights can draft text from spreadsheets. Computer-based managers are being trialled. These allocate work and schedules, with the experience well received by teams of workers to date (Lorentz et al., 2015[5]). Recent software can interpret some human emotion better than humans, presaging new forms of machine-human interaction (OECD, 2017[2]).
The development of new digital technologies powered by Artificial Intelligence (AI, henceforth) occupies a particularly important space in the policy debate that focuses on the impact that digital technologies will have on labour markets. An increasing number of scholars – see for instance (Aghion, Jones and Jones, 2017[6]; Brynjolfsson, Rock and Syverson, 2017[7]; Fossen and Sorgner, 2019[8]) – have adopted the view that AI should be regarded as a General Purpose Technology (GPT) whose fundamental characteristic is that of being able to improve (and self-improve) over time, solving complex problems and generating complementary innovations with little or no human supervision.
In the workplace, the adoption of digital technologies is contributing to a new wave of AI-powered automation. Algorithms are taking on more and more routine tasks, displacing workers from some traditionally cognitive jobs. Muro, Whiton and Maxim (2019[9]) argue for instance that AI is making significant progress in replicating a particular aspect of intelligence, namely “prediction”, this latter being central to decision making and an essential aspect of high-skilled jobs in the health care or business sector. Similarly, (Felten, Raj and Seamans, 2019[10]) and (Webb, 2019[11]), stressed AI’s ability to perform non-routine cognitive tasks through their ability to autonomously “acquire” and “apply” knowledge in problem solving contexts.
New examples, showing the ability of AI to perform such tasks, are being developed at a fast pace. GPT-3 is, for instance, one of the most sophisticated AI-powered Natural Language Processing (NLP) algorithm to this date. The current version of GPT-3 is able to answer complex medical questions and to identify correctly a disease from the simple description of its underlying symptoms, even suggesting the necessary treatment for the disease at hand. Notably, GPT-3 capabilities are transversal, ranging from its ability to write new software code to that of programming mobile applications or to produce autonomously poems and journal articles when prompted with a few lines of text.
Digital technologies played a fundamental role also during the recent COVID-19 pandemic. Evidence (OECD, 2021[12]) shows that the rapid adoption of digital technologies during the coronavirus pandemic has helped protect the jobs of millions of workers who were able to carry out their activities remotely and working from home. The COVID-19 crisis obliged many firms to rethink the way they were doing business and to adjust to a “new normal” where little or no physical interaction was allowed both among co-workers as well as with customers.
As the availability of human manual labour decreased dramatically due to social distancing measures, many firms started adopting new and automated ways to connect with customers. Cashierless stores in the retail sector are a clear example of this trend. Amazon’s checkout-free shopping experience, for instance, leverages a similar technology as the one used in self-driving cars (computer vision, sensor fusion, and deep learning) to automatically detect when products are taken from or returned to the shelves, keeping track of them in a virtual cart and allowing customers to exit the shop without having to queue at the counter before leaving, eliminating the need for cashiers or self-service checkout assistants. Sainsbury’s is offering a similar service through its SmartShop system (Wallace-Stephens and Morgante, 2020[13]) which allows customers to scan their groceries as they go around the store and pay via an app. Sales from this service have reportedly increased from 15% to 30%.
While some of these new technologies were developed prior to the pandemic, the speed by which they have been adopted recently has increased considerably. To give an example, reports in the United Kingdom from the ONS Business Impact of Coronavirus Survey (BICS), indicate that almost one in three businesses have increased their use of online services to help communicate with customers since the pandemic and Sainsbury’s former CEO suggested that the use of digital technologies to eliminate cashiers “might have taken three or four years to get to, [but] it happened in the space of less than six weeks” during the pandemic.
While many employers and employees have used digital technologies to weather the COVID-19 crisis, in particular by adopting teleworking arrangements, others have instead been unable to do so due to the lack of adequate skills or the necessary technological infrastructures in their workplace. This is one of the many examples of how the digital transition could create, or further widen, divides and inequality.
The increasing ability of digital technologies to perform routine as well as cognitive tasks in a manner that is as effective as humans will have a large impact on the way services are delivered, products are manufactured and innovation itself is created (Autor, Levy and Murnane, 2003[14]). Technology can replace workers in routine tasks that are easy to automate and complement workers in tasks that require creativity, problem solving and cognitive skills. As machine learning and artificial intelligence advance in many sectors, a growing number of workers may need to move from declining occupations (which are highly intensive in low-skilled routine tasks) to growing ones (which are characterised by high-level, non-routine cognitive skills).
The fast-paced adoption of digitally powered technologies will certainly have fundamental repercussions on the type of skills that individuals will need to master in at least two separate ways. On the one hand, individuals will need to develop adequate digital and cognitive skills to interact with digital technologies. Digital tools, in fact, do not operate in a vacuum and much of their potential is determined by how well workers are able to interact with them. This is for instance the case of AI, where workers are expected to supply the correct inputs to AI-algorithms and, more importantly, to critically understand the outputs that are produced in return. In other words, individuals will need to be educated on how to detect biases, fakes and mistakes that could result from the misuse of AI.2
On the other hand, digital technologies are likely to replace humans in specific cognitive tasks at work, freeing up time of human labour to perform other tasks that AI is still not capable of doing effectively. Socio-emotional skills and all those traits that make us “humans” (i.e. empathy, intuition and creativity) are expected to become increasingly more important in future labour markets as AI is adopted more broadly in society and at work.
When focusing on the impact of technology on labour markets and jobs, recent research (OECD, 2019[1]) suggests that the adoption of traditional automation technologies (for instance, industrial robotics applied to narrow and repetitive tasks) has, indeed, contributed to the polarisation of skill demands where occupations using mid-level skills have been those most affected by automation due to the routine nature of their tasks.
Technologies can potentially complement and amplify human potential in combinatorial ways. Going forward, the overall impact of digitalisation on jobs will therefore depend on how much digital technologies will complement rather than substitute workers in specific tasks. Today, for example, advances in software and data science help to develop new materials and discover new drugs. In turn, new materials might replace silicon semiconductors with better-performing substrates, allowing more powerful software applications and new drugs can open up the way for novel treatments.
Despite initial fears of potential massive technological unemployment, recent evidence suggests that employment levels have been trending upwards, with the exception of the period of global financial crisis (GFC) and the recent COVID-19 pandemic. Techno-optimists argue that new digital and other technologies will raise productivity (Brynjolfsson and Mcafee, 2014[15]), and that economic history provides reasons to think that technological progress could even accelerate (Mokyr, Vickers and Ziebarth, 2015[16]). A further argument of techno-optimists is that official measures of economic growth understate progress, because they poorly capture many of the benefits of new goods and services. For example, national statistical offices usually collect no information on the use of mobile applications, or online tax preparation, or business spending on databases while the consumer surplus created by hundreds of new digital products is absent from official data (Mandel, 2012[17]).
Other studies, however, point to significant potential labour market challenges. Acemoglu and Restrepo (2020[18]) argue, for instance, that recent declines in the share of labour in national income and the employment to population ratio in the United States – e.g. Karabarbounis and Neiman (2013[19]) and Oberfield and Raval (2014[20]) support the claims that, as digital technologies, robotics and artificial intelligence penetrate the production workflows of many countries, workers will find it increasingly difficult to compete against machines, and their compensation will experience a relative or even absolute decline.
Regardless of whether the digital transition will create net employment gains, anxieties remain as to the already existing inequalities in the access and use of technology may increase in the near future if effective policy intervention does not support the most vulnerable in developing adequate digital skills.
Individuals lacking adequate skills to use new technologies are, in fact, likely to lag behind and face barriers to engage in a new digital society or in labour markets that increasingly require digital skills.
Recent empirical evidence (Fossen and Sorgner, 2019[8]) suggests that the effects of new digital technologies on employment stability and wage growth are observable at the individual level. In particular, results for the United States in between 2011 and 2018 suggest that high computerisation risk in jobs is associated with a corresponding high likelihood of switching from one occupation to another or of becoming non-employed, as well as to a decrease in wage growth.
The negative impacts are skewed towards the most vulnerable, as results point to highly educated individuals being more able than workers with lower levels of education to adapt to computerisation risk as highly skilled workers are better equipped with skills that cannot be easily automated, such as creative and social intelligence, reasoning skills, and critical thinking.
Given the current speed of technological advancement, hardship could affect workers in specific sectors in an uneven manner. The technology of driverless vehicles is a frequently commented example of such potential displacement. Taken together, just over 3 million people work as commercial drivers in 15 European Union member states. Eliminating the need for drivers could create an exceptional labour market shock (OECD, 2017[2]).
From a policy making standpoint, it is very difficult to precisely predict how new technologies might transform existing specific jobs and as such, the extent of potential raising inequalities. In the banking sector, for instance, it was long believed that automated teller machines (ATMs) would cancel the need for human tellers. While ATMs were introduced in the 1970s, in between 1971 and 1997 the share of human tellers among all workers in US banking only declined modestly, from just under 21% to around 18% (Handel, 2012[21]). Similarly, many technologies, such as big data and the IoT, have developed in a wave-like pattern, with periods of rapid inventive activity coming after periods of slower activity and vice versa (OECD, 2015[22]). Cloud computing, for example, was first commercialised in the 1990s, but by 2017 it had still only been adopted by less than one in four businesses in OECD countries (OECD, 2017[2]).
In such uncertain and rapidly changing scenario, the analysis of timely and granular information on labour market and skill demands is key to understand the extent of the impact of the digital transformation on labour markets and workers.
This report aims to unveil the most relevant labour market trends related to the adoption of digital technologies. It does so by analysing the demand for digital skills and occupations using the detailed information contained in millions of job postings that have been published online by firms and employers across 10 different countries (Belgium, Canada, France, Italy, Germany, the Netherlands, Singapore, Spain, the United Kingdom and the United States) in between the year 2012 and now.
The analysis of online job postings in this report allows to identify the major shifts in the demand for digital skills with an unprecedented level of granularity and to track the evolution of occupational demand over time up to very recent months. The report identifies the skill profiles of the most relevant digital occupations, by highlighting the emerging adoption of new technologies, tools and digital skills across a variety of job roles.
Similarly, the report uses the detailed skill information contained in online job postings to identify skill complementarities across a variety of different occupations in order to suggest potential career transitions from occupations that are in decline to others that are thriving in the digital labour market.
References
[18] Acemoglu, D. and P. Restrepo (2020), “Robots and Jobs: Evidence from US Labor Markets”, Journal of Political Economy, Vol. 128/6, pp. 2188-2244, https://doi.org/10.1086/705716.
[6] Aghion, P., B. Jones and C. Jones (2017), Artificial Intelligence and Economic Growth, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w23928.
[14] Autor, D., F. Levy and R. Murnane (2003), “The Skill Content of Recent Technological Change: An Empirical Exploration”, The Quarterly Journal of Economics, Vol. 118/4, pp. 1279-1333, https://doi.org/10.1162/003355303322552801.
[15] Brynjolfsson, E. and A. Mcafee (2014), The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies, W.W. Nortan & Company Inc.
[7] Brynjolfsson, E., D. Rock and C. Syverson (2017), Artificial Intelligence and the Modern Productivity Paradox: A Clash of Expectations and Statistics, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w24001.
[10] Felten, E., M. Raj and R. Seamans (2019), “The Variable Impact of Artificial Intelligence on Labor: The Role of Complementary Skills and Technologies”, SSRN Electronic Journal, https://doi.org/10.2139/ssrn.3368605.
[8] Fossen, F. and A. Sorgner (2019), “New Digital Technologies and Heterogeneous Employment and Wage Dynamics in the United States: Evidence from Individual-Level Data”, IZA Discussion Papers 12242, https://www.iza.org/publications/dp/12242/new-digital-technologies-and-heterogeneous-employment-and-wage-dynamics-in-the-united-states-evidence-from-individual-level-data.
[21] Handel, M. (2012), “Trends in Job Skill Demands in OECD Countries”, OECD Social, Employment and Migration Working Papers, No. 143, OECD Publishing, Paris, https://doi.org/10.1787/5k8zk8pcq6td-en.
[3] Hoos, H. (2012), “Programming by optimization”, Communications of the ACM, Vol. 55/2, pp. 70-80, https://doi.org/10.1145/2076450.2076469.
[19] Karabarbounis, L. and B. Neiman (2013), “The Global Decline of the Labor Share*”, The Quarterly Journal of Economics, Vol. 129/1, pp. 61-103, https://doi.org/10.1093/qje/qjt032.
[5] Lorentz et al. (2015), “Man and machine in Industry 4.0: How will technology transform the industrial workforce through 2025?, https://www.bcg.com/publications/2015/technology-business-transformation-engineered-products-infrastructure-man-machine-industry-4.
[17] Mandel, M. (2012), Beyond goods and services: The (unmeasured) rise of the data-driven economy, Progressive Policy Institutue, http://www.progressivepolicy.org/2012/10/beyond-goods-and-services-theunmeasured-rise-of-the-data-driven-economy/.
[16] Mokyr, J., C. Vickers and N. Ziebarth (2015), “The History of Technological Anxiety and the Future of Economic Growth: Is This Time Different?”, Journal of Economic Perspectives, Vol. 29/3, pp. 31-50, https://doi.org/10.1257/jep.29.3.31.
[9] Muro, M., J. Whiton and R. Maxim (2019), Brookings metropolitan policy program.
[20] Oberfield, E. and D. Raval (2014), Micro Data and Macro Technology, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w20452.
[12] OECD (2021), OECD Skills Outlook 2021: Learning for Life, OECD Publishing, Paris, https://doi.org/10.1787/0ae365b4-en.
[1] OECD (2019), OECD Employment Outlook 2019: The Future of Work, OECD Publishing, Paris, https://doi.org/10.1787/9ee00155-en.
[2] OECD (2017), The Next Production Revolution: Implications for Governments and Business, OECD Publishing, Paris, https://doi.org/10.1787/9789264271036-en.
[22] OECD (2015), Data-Driven Innovation: Big Data for Growth and Well-Being, OECD Publishing, Paris, https://doi.org/10.1787/9789264229358-en.
[4] Rotman, D. (2013), “How Technology Is Destroying Jobs”, MIT Technology Review June, http://www.technologyreview.com/s/515926/how-technology-is-destroying-jobs/.
[13] Wallace-Stephens, F. and E. Morgante (2020), Who is at Risk?: Work and automation in the time of Covid-19, https://www.thersa.org/reports/work-automation-covid.
[11] Webb, M. (2019), “The Impact of Artificial Intelligence on the Labor Market”, SSRN Electronic Journal, https://doi.org/10.2139/ssrn.3482150.
Notes
← 1. In using the term Industry 4.0, the contrast is made with three previous industrial revolutions. These three revolutions can only be dated approximately. They are: i) the advent of steam-powered mechanical production equipment (1780s, or thereabouts); ii) electrically powered mass production (1870s); and iii) electronically based, automated production (1960s) (although with many differences compared to the electronics of Industry 4.0, e.g. in terms of cost, size, computational power, intelligence, interconnectivity, and integration with material objects) (OECD, 2017[2]).
← 2. AI has been recently used to produce so-called “deep fakes”, that is videos posted online where celebrities as well as politicians would be seen acting or saying things that never happened in reality. The degree of sophistication of these deep fakes makes them, in many cases, indistinguishable from real videos and it poses the fundamental question of educating people to detect them. Similarly, NLP algorithms such as GPT-3 have been criticised for producing content that could be gender or race biased as the text used to train the algorithm was directly downloaded from the internet without any filter and where these biases may be present in blogs and media content.