Dear all, Upon inheriting the Looking Glass from our predecessors, we identified a number of key issues. Firstly, there were simply not enough articles being published, due both to a lack of submissions from the school community and limited responsiveness from the previous Academic Team. Secondly, the Looking Glass had not been advertised or explained effectively enough to the wider school community. As a result, we plan to implement a more consistent and engaging stream of articles on the Looking Glass. As part of this initiative, we are looking to recruit a select group of keen writers from across the lower school who would be willing to produce one high-quality piece of writing, discussion, or media each month for publication on the Looking Glass. We believe this will be hugely beneficial both to the school community, which will gain access to a wider range of opinions and viewpoints, and to prospective writers, who will be able to reference their experience contributing to the Look...
Economics: Will Artificial Intelligence trigger the next productivity revolution in developed economies?
Note: The following article was written by Siddharth Sethi L6 (20SethiS@students.watfordboys.org)
Growth led by productivity has long been seen as the driving force to increase living standards. Paul Krugman claimed that ‘productivity isn’t everything, but in the long run it is almost everything’. Since the early 21st century, many developed economies have experienced weak productivity, especially the UK, giving rise to concerns about stagnating wages and improvements to living standards.
Figure 1. GDP per hour worked in
US$ within selected developed
economies, 1990-2023. Source:
The Productivity Institute.
Although AI has significant potential to trigger a new productivity revolution, historical experience suggests that technological breakthroughs alone are insufficient. Gains will ultimately depend on complementary investments in human capital and infrastructure, and if institutions ensure that the benefits are shared rather than concentrated in the wealthy.
One of the strongest arguments for AI generating a productivity revolution is due to its potential role as a general purpose technology. This will reshape production virtually in all sectors of the economy. The framework of the Solow Growth model states that any long run increase in output per worker is driven primarily by improvements in total factor productivity. AI directly and efficiently contributes to this process by automating routine cognitive tasks and optimising the way both labour and capital are utilised. The endogenous growth theory developed by Paul Romer, shows that AI is significant because it can accelerate the production of ideas itself. An empirical study by researchers at MIT and Stanford examined well over 5000 customer support agents at a Fortune 500 company in 2023. Within this study it was discovered that access to a generative AI assistant increased productivity by approximately 14%. The gains were even more substantial for less experienced workers with it exceeding 34%. The researchers concluded that AI ‘disseminates the best practices of more able workers and helps newer workers move down the experience curve’. Similarly, a controlled experiment conducted by the cloud based developer platform ‘Github’ found that software developers using the Github Copilot feature completed programming tasks 55.8% faster than those without access to the AI. These findings within these academic research papers suggest that AI can significantly increase output per worker, especially in knowledge intensive industries. AI does not merely complement existing production methods. Instead it
could be considered to have the potential to fundamentally transform them, creating the conditions necessary for rapid but most importantly sustained increase in productivity driven growth.
Despite AI’s considerable potential, history suggests that transformative technologies rarely translate into immediate productivity gains. This is encapsulated by the Solow Productivity Paradox. Robert Solow, Nobel Laureate, stated in 1987 that ‘you can see the computer age everywhere but in the productivity statistics’. Solow’s paradox highlights that while technological innovation may advance rapidly, aggregate productivity often increases over a long period. This is typically due to the time required for firms to reorganise their production and retrain workers as well as invest within complementary capital. A historic example of this is electricity. Although commercially introduced during the late 19th century, substantial productivity gains did not emerge until the 1920s. This was due to the time it takes to redesign factories to be able to exploit electric powered assembly lines rather than simply switching from steam engines to electric motors. Despite the widespread computer adoption in the 1980s, labour productivity only increased from the mid to lte 1990s. Building on this historical perspective, economist Robert Gordon argued that AI may not replicate the transformative impact of general purpose technologies such as electricity=. Instead he contend that AI is more likely to improve existing processes rather than revolutionise them. He implies that its macroeconomic effects may be more incremental than many experts forecast. Hence, while AI undoubtedly possesses significant productivity enhancing potential, it would be imprudent to assume that an increased technological capability alone will generate a sustained productivity revolution.
While AI has the potential to raise productivity in the long run its widespread adoption is also likely to accelerate labour market disruption. Economic theory suggests that technological progress is often characterised by creative destruction. This process was identified by Schumpeter where innovation simultaneously creates new industries but also rendering existing jobs and industries obsolete. AI is effective at automating routine cognitive tasks, putting administrative support, customer service, and basic research at risk of displacement. In 2023, Goldman Sachs estimated that generative AI could expose the equivalent of 300 million full time jobs worldwide to automation. Similarly the International Monetary Fund estimated that 40% of jobs globally, and 60% in developed economies could be affected by AI, reducing income for many households. However history suggests that technological revolutions have ultimately generated more employment than it has destroyed. The Industrial Revolution displaced millions of workers in the short run, but simultaneously created entirely new industries, with the UK focusing on providing services. AI also will create new occupations such as data scientists, machine learning engineers as well as AI safety specialists. The extent to which AI triggers a productivity revolution ultimately falls upon governments. Without such measures AI may be able to increase productivity, however it can exacerbate unemployment and income inequality.
Artificial intelligence has all the characteristics of a general purpose technology. AI can do this by increasing innovation and efficiency of both labour and capital. AI can reverse the productivity slowdown that has restricted growth in many economies in the past 2 decades. However history demonstrates that technological breakthroughs alone are rarely sufficient to transform economic performance. The introduction of electricity and computers illustrate that productivity gains emerge if complementary investments occur in infrastructure. The
displacement of workers and the risk of widening inequality demonstrates how higher productivity does not translate into higher living standards. Whether AI triggers a productivity revolution will depend less on the technology itself and more on the policies that shape its adoption. Governments that invest in education and support innovation will maximise AI’s economic benefits. Those who fail to adapt risk allowing one of the most transformative technologies of the 21st century to deepen existing structural challenges rather than resolve them.
References
Brynjolfsson, E., Li, D. and Raymond, L. (2023) Generative AI at Work. National Bureau of Economic Research Working Paper No. 31161. Available at:
https://www.nber.org/papers/w31161
Peng, S., Kalliamvakou, E., Cihon, P. and Demirer, M. (2023) The Impact of AI on Developer Productivity: Evidence from GitHub Copilot. Available at: https://arxiv.org/pdf/2302.06590
Brynjolfsson, E. and Hitt, L.M. (2000) Beyond Computation: Information Technology, Organizational Transformation and Business Performance. Journal of Economic Perspectives. Available at:
https://faculty.wharton.upenn.edu/wp-content/uploads/2012/04/Beyond-computation.pdf
David, P.A. (1990) The Dynamo and the Computer: An Historical Perspective on the Modern Productivity Paradox. American Economic Review. Available At:
https://www.almendron.com/tribuna/wp-content/uploads/2018/03/the-dynamo-and-the-comp uter-an-historical-perspective-on-the-modern-productivity-paradox.pdf
Briggs, J. and Kodnani, D. (2023) The Potentially Large Effects of Artificial Intelligence on Economic Growth. Goldman Sachs Global Investment Research.
https://blog.biocomm.ai/2023/03/26/goldman-sachs-global-economics-analyst-the-potentially -large-effects-of-artificial-intelligence-on-economic-growth-briggs-kodnani-26-march-2023/
Comments
Post a Comment