Will Robots Take Your Job? Automation, Work, and the Political Economy of Technology with Ha-Joon Chang

This podcast episode features Ha-Joon Chang discussing the historical and contemporary impacts of automation, including robots and artificial intelligence, on the world of production and future working life. He challenges the common narrative that automation inevitably leads to widespread job destruction, emphasizing that technological possibilities are only one part of a complex story. Chang traces the history of automation from the 18th century to modern CNC machines and industrial robots, highlighting that fears of job displacement are not new, as seen with the Luddites and Keynes's concept of 'technological unemployment'. He critiques neoclassical economic theory for often treating technology as an exogenous factor, arguing instead for a more nuanced understanding of technological development as a lengthy process of investment, adaptation, learning, and capability accumulation, which is deeply intertwined with economic and political decisions.

The episode delves into two main concerns raised by non-neoclassical economists: the worry about unemployment and the dehumanization of work. Chang cites studies from the OECD and McKinsey Global Institute, which suggest that a relatively low percentage of jobs are at high risk of full automation, with most jobs experiencing task reconfiguration rather than complete replacement. He introduces the concept of 'asset specificity' by Oliver Williamson to explain why expensive, specialized automation technologies are not always adopted enthusiastically, especially during periods of economic uncertainty. Furthermore, he discusses 'dynamic demand responses,' where increased productivity and falling prices can lead to a greater overall demand for goods and services, potentially creating more jobs than initially lost, and 'dynamic supply responses,' which generate new jobs for maintaining and developing new technologies.

Chang argues that the current era, characterized by 'digital factories' and advancements in 'machine learning,' might represent a qualitative shift, as machines begin to simulate human learning and affect white-collar jobs. However, he also points out 'Moravec's Paradox,' where robots find difficult things easy and easy things difficult, suggesting limitations to current AI. Crucially, the podcast emphasizes the political dimension of technological change. Chang, drawing on Karl Marx and other Marxist economists like Harry Braverman and Stephen Marglin, contends that capitalists often choose technologies that maximize control over workers, not necessarily the most efficient ones. He highlights how government regulations, subsidies, and societal choices can deliberately shape the evolution and adoption of technology, for instance, by promoting robots for elder care rather than car manufacturing.

Finally, the episode explores how government policies regarding worker training, retraining, and social welfare significantly influence the level of technological unemployment. Countries with robust welfare states and employment guarantees, like Sweden and Japan, have seen less worker resistance to automation compared to the United States. Chang concludes that the decision to embrace or slow down automation is a philosophical and political choice, balancing material comfort and leisure against the human need for work, identity, and social interaction. He stresses that assuring workers about their livelihood and future is key to facilitating technological adoption without causing widespread misery or resistance.

"almost 50% of the jobs in the u.s. may be automated over the next decade or two"

"industrial robots are defined as automatically control reprogrammable multi-purpose manipulator programmable in three or more axes which may be either fixed in place or mobile for using industry automation applications"

"you can't just impose some technology and use it like that I mean you need a whole process of adaptation and learning and the acquisition of ability abilities to use and improve but the input the technology which is known as technological capabilities"

"worry about what Keynes called ecological employment if you use more machines this will lead to redundancy of people this will create unemployment"

"machines would lead to the dehumanization workers"

"less than actually five percent of Occupation can be fully automated"

"the technological possibilities for automation are only part of the story first of all the fact that something can be automated doesn't lessen that it will be automated"

"robots find difficult things easy and easy things difficult so this is are known as moriches paradox"

"to the capitalist human workers are trouble yeah but the old mind their free will they can get organized you know the okona strike so if they can get away with it they want to get rid of all human workers"

"society without us necessarily realizing is doing lots of things through government regulations subsidies bans shape the evolution of Technology"

"a large part of our human identity and capabilities are significantly shaped by our work life"

"whether automation technology gets introduced or not is you know partly affected by things like the welfare state and employment guarantees and some because workers yeah they cannot resist it forever but they can for a significant period resist introduction of certain automation technologies"