By Maarten Goos
During the US-EU Trade and Technology Council (TTC) in late September 2021 in Pittsburgh, both the US and European Commission expressed strong interest in working on a joint study to assess the potential impact of Artificial Intelligence on our workforces. The Pittsburgh Statement committed to a joint “economic study examining the impact of AI on the future of our workforces, with attention to outcomes in employment, wages, and the dispersion of labor market opportunities.” Through this collaborative effort, the TTC intends to inform approaches to AI consistent with an inclusive economic policy that ensures the benefits of technological gains are broadly shared by workers across the wage scale.
On December 5, 2022 the TTC met outside Washington, D.C., and presented a final version of the report written by TPRI associate Maarten Goos representing the European Commission, and Michael Sinkinson, John Iselin, and Lindsey Raymond from the Council of Economic Advisors. The report summarizes the most recent insights into the economic impact of AI with a focus on labor markets.
The first part of the report defines AI and discusses its overall potential benefits and harms. The type of AI that has driven recent excitement is Machine Learning, and Deep Learning based on neural networks in particular. Machine Learning is a branch of computational statistics that focuses on designing algorithms to make predictions from data without explicitly programming the solution. Since 2012, the uses of Machine Learning as a prediction technology have grown substantially and have become ubiquitous: Pandora learns how to make better music recommendations based on its users’ preferences; Google learns how to automatically translate content into different languages based on translated documents found online; and Facebook learns how to identify people in photos based on its database of known users.
As AI continues to evolve and find its way into a wide variety of applications, it will generate economic benefits. However, AI also poses several challenges such as privacy violation, unfair competition through unequal access to data, behavioral manipulation by algorithms, the impact of AI on communication, echo chambers in social media, or the ability of governments to closely monitor dissent through AI. The report further explains that like earlier technologies, AI’s potential cannot be realized without a proper understanding and management of these challenges.
The second part of the report summarizes recent studies that discuss the adoption of AI by firms. In both the US and EU, only a small fraction of firms reports to use AI. In contrast to robotics that is mainly used in manufacturing, AI is being adopted in all sectors of the economy. Within sectors, larger and younger firms are more likely to adopt AI. The fact that AI adoption concentrates in larger and younger firms reflects that there are substantial costs and organizational barriers involved in adopting AI. In the US, firms report the inapplicability of AI to its business and AI being too costly as the main reasons for not adopting AI. In Europe, skills are also an important barrier reported by firms, with about 80% of enterprises citing a lack of skills in their internal workforce and in the external labour market.
The third part of the report focuses on the impact of AI on work. AI is likely to fundamentally change workplaces and labor relations. In contrast to previous episodes of digital breakthroughs, AI can infer tacit knowledge that need not be fully specified by underlying software because it learns to perform tasks inductively by training on examples instead of by following explicit rules that are programmable.
One important question that the report focuses on is what tasks done by workers will be automated by AI? While earlier digital technologies automated occupations intensive in doing routine tasks (e.g., machine operators, office clerks), machine learning as a prediction technology has the potential of also automating various non-routine tasks across a wide range of occupations. Most-exposed occupations include clinical laboratory technicians, chemical engineers, optometrists, and power plant operators. More generally, high-skill occupations are most exposed to AI. There is also a small proportion of low-skilled jobs that are highly exposed to AI. Examples are production jobs that involve inspection and quality control.
Another important question that is discussed in the report is the impact of algorithmic management on labor relations. Algorithmic management relies on data collection and surveillance of workers to manage workforces in an automated way. Online labor platforms are a well-known example. These platforms enable workers to choose the clients and jobs they take, how they carry out those jobs, and the rates they charge to do them. However, to varying degrees, workers’ ability to make these choices is strongly shaped by platform rules and design features. Increasingly, algorithmic management is also being used in other settings, such as in warehouses, retail, manufacturing, marketing, consultancy, banking, hotels, call centers, and among journalists, lawyers, and the police. The report illustrates this further with case studies about algorithmic hiring and algorithmic management in warehouses. The report also discusses the broader impact that algorithmic management has on business models, and how this results in fissured labor relations.
Finally, the report ends with some conclusions for policy makers such as encouragement of development and adoption of AI that is beneficial for workers or investing in training and job transition services so that those workers most disrupted by AI can move to new positions where their skills and experience are most applicable. Importantly, governments should invest more in the capacity of regulatory agencies to ensure that AI systems are transparent and fair for workers. Governments are drafting blueprints to move in that direction. The US has announced an initiative to create a “Bill of Rights” for AI covering many areas, such as consumer protections and equity of opportunity in employment, education, housing and finance, and health care. A similar initiative exists in Europe with the European Commission Artificial Intelligence Act. Only if these blueprints result in regulation that governs AI properly and timely will AI lead to economic growth, shared prosperity, and substantially greater welfare for society.