ssrn-6154633

Karan Aggarwal, East Asia, Southeast Asia, and Oceania, Prompt 1 Artificial Intelligence, Economic Progress, and Uneven Adjustment Artificial intelligence has become one of the most influential sources of technological change in modern economies. Its economic importance lies not only in its ability to automate tasks, but also in how it reshapes productivity, employment, and the distribution of resources across society. This essay argues that artificial intelligence is a major driver of technological progress while also increasing the risk of uneven economic adjustment, particularly during periods of economic slowdown. By changing the relative cost of tasks and enabling productivity gains with limited capital, artificial intelligence can accelerate growth while shifting resources toward some groups and away from others when labor markets do not adjust smoothly. Technological Progress and Task Costs From an economic perspective, artificial intelligence affects production mainly by altering the relative prices of tasks rather than eliminating entire occupations. Tasks based on information processing, such as preparing presentations, tracking sales, or performing routine accounting, have become significantly cheaper when carried out using artificial intelligence systems. These tasks require little physical capital and can be scaled quickly once software infrastructure is established. In contrast, tasks that involve physical interaction with the environment, including logistics management, food preparation, and shipping, often require expensive hardware, specialized infrastructure, and large amounts of training data. As a result, automation in these areas can be costly or technologically constrained. This difference in task costs helps explain why artificial intelligence adoption has been faster in white collar and analytical functions than in many physical production 1 -- 1 of 5 -- Karan Aggarwal, East Asia, Southeast Asia, and Oceania, Prompt 1 processes. Artificial intelligence is also not yet reliable across all contexts, which means continued investment in training, supervision, and system refinement remains necessary. As a result, progress driven by artificial intelligence is uneven across sectors and depends heavily on feasibility and economic cost rather than technological possibility alone. Productivity Gains with Low Capital Requirements Traditional growth models link productivity improvements to substantial capital accumulation. Artificial intelligence challenges this relationship. When deployed at scale, artificial intelligence can generate large productivity gains with relatively low marginal capital investment. For example, evidence from the OECD shows that firms in digitally advanced Asian economies have achieved significant efficiency improvements through software based automation without comparable increases in physical capital expenditure. Once trained, artificial intelligence systems can be replicated at near zero cost, allowing firms to raise output or efficiency without proportional increases in traditional inputs. This dynamic is particularly important for small firms and startups. With access to artificial intelligence tools, firms with limited capital can reach productivity levels that were previously associated with much larger organizations. At the same time, these gains depend on access to data, skills, and complementary organizational practices. Firms that fail to integrate artificial intelligence effectively may see little improvement, leading to growing productivity gaps within industries rather than broad based gains. Employment, Recession, and Uneven Adjustment Although artificial intelligence supports long term economic progress, its interaction with labor markets during downturns raises important concerns. Artificial intelligence tends 2 -- 2 of 5 -- Karan Aggarwal, East Asia, Southeast Asia, and Oceania, Prompt 1 to reduce demand for routine tasks, and during periods of weak growth or recession, displaced workers may struggle to transition into complementary roles. This can cause unemployment to rise in a more persistent and uneven way, as task displacement occurs faster than job creation. Research by David Deming emphasizes the increasing importance of social and analytical skills that complement technology, suggesting that workers without these skills are more vulnerable during technological transitions. When retraining systems and job matching mechanisms function poorly, income and opportunities tend to shift toward groups that are already better positioned to benefit from artificial intelligence adoption. While such outcomes are not inevitable, they are more likely when the pace of technological change exceeds the capacity of institutions to adjust. A common counterargument is that technological change eventually creates more jobs than it destroys, since productivity gains lower prices and stimulate demand in other parts of the economy. While this has often been true over long historical periods, it depends on relatively smooth labor reallocation. In the short to medium run, especially during recessions, task displacement can outpace the creation of complementary roles, resulting in persistent unemployment for certain groups. From an economic standpoint, the central issue is therefore not net job loss, but the speed and unevenness of adjustment. Firms, Strategy, and Organizational Change Artificial intelligence also affects how firms are organized and how strategic decisions are made. Research by Anita Elberse highlights that data and analytics improve firm performance when combined with effective organizational choices rather than replacing 3 -- 3 of 5 -- Karan Aggarwal, East Asia, Southeast Asia, and Oceania, Prompt 1 them. By lowering information and coordination costs, artificial intelligence improves decision making, but it does not remove the need for human judgment, leadership, or oversight. Firms that successfully integrate artificial intelligence with strong management practices gain a competitive advantage, while others may fall behind. This can contribute to greater concentration within markets. At the same time, artificial intelligence does not automatically favor large firms. By reducing coordination costs, it can also allow smaller firms to scale decision making more efficiently, potentially reshaping competitive dynamics rather than reinforcing a single market structure. Conclusion: Progress with Trade Offs Artificial intelligence is an important source of economic progress because it enables large productivity gains with relatively low capital requirements and reduces the cost of many cognitive tasks. These features support long term growth and innovation across a wide range of sectors. However, the same mechanisms that drive progress can also intensify adjustment challenges, particularly during economic downturns. Uneven task displacement, skill biased effects, and slow labor market adjustment can shift resources toward some groups while disadvantaging others. From an ethical perspective, the tension arises because productivity gains are measured at the aggregate level, while adjustment costs are borne by specific workers and regions. Understanding the economic impact of artificial intelligence therefore requires recognizing both its transformative potential and the risks associated with uneven adaptation. 4 -- 4 of 5 -- Karan Aggarwal, East Asia, Southeast Asia, and Oceania, Prompt 1 Works Cited Deming, David J. “The Growing Importance of Social Skills in the Labor Market.” Quarterly Journal of Economics, vol. 132, no. 4, 2017, pp. 1593–1640. Elberse, Anita. Blockbusters: Hit Making, Risk Taking, and the Big Business of Entertainment. Henry Holt and Company, 2013. Solow, Robert M. “Technical Change and the Aggregate Production Function.” The Review of Economics and Statistics, vol. 39, no. 3, 1957, pp. 312–320. Organisation for Economic Co-operation and Development. OECD Employment Outlook 2023: Artificial Intelligence and the Labour Market. OECD Publishing, 2023. Kapoor, Parikshit. Personal communication. Discussion on artificial intelligence and economic adaptation, 2025. 5 -- 5 of 5 --