By Francesco Filippucci, Peter Gal, Cecilia Jona-Lasinio, Alvaro Leandro, Giuseppe Nicoletti
Recent OECD work discusses the impact of Artificial Intelligence (AI) on productivity, distribution, and growth, highlighting the challenges and conditions to be met before its benefits can be realised (Filippucci et al., 2024). These issues are particularly timely considering weak productivity growth across OECD economies in recent decades (Andre and Gal, 2024) and the widespread enthusiasm and active debate surrounding the growth-potential of AI. Some proponents suggest it might reverse the long-standing productivity slowdown by adding 1-1.5 percentage points of annual growth (Baily, Brynjolfsson, and Korinek, 2023; Artificial Intelligence Commission of France, 2024; Briggs and Kodnani, 2023). Acemoglu (2024), on the other hand, contends that current capabilities can only support moderate macro-level productivity gains, in the order of 0.1% per year.
AI is becoming a general-purpose technology with a transformative impact on a broad range of economic activities, as was the case with computers, the internet or electricity (Agrawal, Gans and Goldfarb, 2019). It is a digital technology that combines software, data and computing power to perform a variety of advanced cognitive tasks, such as content generation, prediction, or even physical tasks (coupled with robotics). It is also characterised by self-improvement (learning) and greater autonomy. These features imply that AI can boost not only the production of goods and services but also the generation of ideas, speeding up research and innovation (Aghion, Jones and Jones, 2018).
Reviewing the fast-growing literature, initial micro-level evidence covering firms, workers and researchers suggests that AI may stimulate innovation (Van Noorden and Perkel, 2023) and delivers significant productivity and performance benefits. The size of the firm-level productivity gains from pre-Generative AI is comparable to previous digital technologies (up to 10%; see Figure 1, panel on Non-Generative AI). When using more recent Generative AI to assist with various tasks – writing, computer programming or customer service – substantially larger performance benefits have been identified, with widely varying magnitudes depending on the context (in the order of 20-50%; see panel on Generative AI).
Figure 1. The relationship between AI and productivity or worker performance: Selected estimates from the literature
Note: *controlling for other ICT technologies. In the Non-Generative AI panel, “AI use” is a 0-1 dummy obtained by firm surveys, while AI patents refers either to a 0-1 dummy for having at least 1 patent (US study) or to the number of patents in firms (for the EU+UK study, where the average number is 0.48 with 2.6 standard deviation, so that firms cumulating more than one patents are relatively few). Two of the estimates in the panel (“9 countries, 2016-21”) relate to the same study (Calvino and Fontanelli, 2023), but the second estimate controls for other ICT technology use and thus better isolates the marginal impact of AI. Given that the study reports separate estimates for all 9 countries, the median estimate across countries is shown on the Figure.
Source: authors’ compilation from micro level studies.
How these microeconomic gains will translate into macroeconomic productivity growth hinges on the extent of AI adoption, which seems limited to date at less than 5% of firms in the United States (Census Bureau, 2024). It also depends on whether AI-driven automation displaces workers from heavily affected activities; or the human-augmenting capabilities of AI will prevail, underpinning labour demand. Currently, AI exposure varies greatly across sectors: knowledge-intensive, high-productivity activities seem much more affected (Figure 2), with high potential for automation in some cases (Cazzaniga et al, 2024; WEF, 2023). Hence an eventual fall in the employment shares of these sectors would act as a drag on aggregate productivity growth, resembling a new form of “Baumol disease” (Aghion, Antonin and Bunel, 2019).
Figure 2. High productivity and knowledge intensive services are most affected by AI
AI exposure of workers by sector (2019)
Note: The index measures the extent to which worker abilities are related to important AI applications. The measure is standardized with mean zero and standard deviation 1 at the occupation level and then matched to sectors. The figure does not yet include recent Generative AI models.
*Including non-market services, manufacturing, utilities, etc.
Source: Filippucci et al (2024) and OECD (2024) based on (Felten, Raj and Seamans, 2021).
AI-driven threats to market competition and inequality may weigh on its potential benefits. First, the high fixed costs and returns to scale related to data and computing power may lead to excessive concentration of AI development. Second, AI use in downstream applications may also lead to market distortions, especially if it allows first movers to build up a substantial lead in market share and market power. Moreover, AI-powered pricing algorithms have a tendency to charge supra-competitive prices (Calvano et al., 2020), and can also enhance harmful price discrimination (OECD, 2018).
AI will likely have ambiguous impacts on inequality. The technology has the potential to substitute for high-skilled labour and narrow wage gaps with low-skilled workers, thereby reducing inequalities (Autor, 2024) at least within occupations (Georgieff, 2024). But there are also indications that AI can be associated with higher unemployment (OECD, 2024). On the other hand, AI can also lead to more inclusion and stronger economic mobility by improving education quality and access, expanding credit availability, and lowering skill barriers (e.g. foreign languages).
Further uncertainties surrounding AI include broader societal concerns. More immediate ones relate to privacy, misinformation, and bias (possibly leading to exclusion), while longer-term ones include mass unemployment or even existential risks (Nordhaus, 2021; Jones, 2023).
A comprehensive policy approach is needed to effectively manage these risks and harness AI’s full potential. Immediate priorities involve promoting market competition and widespread access to AI technologies while preserving innovation incentives and addressing issues of reliability and bias. Job displacement, reallocation and inequality impacts might emerge over longer periods, but they require preventive policy action through training, education, and redistribution measures. Policymakers should also devise national and international governance mechanisms to cope with rapid, unpredictable developments in AI.
Endnote:
The main paper underlying this blog (Filippucci et al, 2024) was developed within the Joint OECD-Italy’s Department of Treasury Project for Multilateral Policy Support.
References:
– Acemoglu, D. (2024), “The Simple Macroeconomics of AI”, https://economics.mit.edu/sites/default/files/2024-04/The%20Simple%20Macroeconomics%20of%20AI.pdf.
– Aghion, P., B. Jones and C. Jones (2018), “Artificial Intelligence and Economic Growth”, NBER Chapter in The Economics of Artificial Intelligence: An Agenda, p. 237-28, https://doi.org/10.3386/w23928.
– Aghion, P., C. Antonin and S. Bunel (2019), “Artificial Intelligence, Growth and Employment: The Role of Policy”, Economie et Statistique / Economics and Statistics, 510-511-512, 149–164. https://doi.org/10.24187/ecostat.2019.510t.1994.
– Agrawal, A., J. Gans and A. Goldfarb (2019), “Economic Policy for Artificial Intelligence”, Innovation Policy and the Economy, Vol. 19, https://doi.org/10.1086/699935.
– Andre, C., and P. Gal (2024), “Reviving productivity growth: A review of policies”, OECD Economic Policy Papers, forthcoming.
– Artificial Intelligence Commission of France (2024), “AI, Our Ambition for France” / “IA : Notre Ambition pour la France”, https://www.info.gouv.fr/upload/media/content/0001/09/4d3cc456dd2f5b9d79ee75feea63b47f10d75158.pdf.
– Autor, D. (2024), “Applying AI to Rebuild Middle Class Jobs”, NBER Woking Paper, No. 32140, National Bureau of Economic Research, https://doi.org/10.3386/w32140.
– Baily, M., E. Brynjolfsson and A. Korinek (2023), “Machines of mind: The case for an AI-powered productivity boom”, in The Economics and Regulation of Artificial Intelligence and Emerging Technologies, Brookings, https://www.brookings.edu/articles/machines-of-mind-the-case-for-an-ai-powered-productivity-boom/.
– Briggs, J., and D. Kodnani, (2023), “The Potentially Large Effects of Artificial Intelligence on Economic Growth”, Global Economics Analyst, Goldman Sachs, New York, https://www.gspublishing.com/content/research/en/reports/2023/10/30/2d567ebf-0e7d-4769-8f01-7c62e894a779.html.
– Calvano, E. et al. (2020), “Artificial intelligence, algorithmic pricing, and collusion”, American Economic Review, Vol. 110/10, p. 3267-3297, https://doi.org/10.1257/aer.20190623.
– Calvino, F. and L. Fontanelli (2023), “A portrait of AI adopters across countries: Firm characteristics, assets’ complementarities and productivity”, OECD Science, Technology and Industry Working Papers, No. 2023/02, OECD Publishing, Paris, https://doi.org/10.1787/0fb79bb9-en.
– Cazzaniga, M. et al. (2024), “Gen-AI: Artificial Intelligence and the Future of Work”, IMF Staff Discussion Notes, International Monetary Fund, https://www.imf.org/en/Publications/Staff-Discussion-Notes/Issues/2024/01/14/Gen-AI-Artificial-Intelligence-and-the-Future-of-Work-542379.
– Census Bureau (2024), Business Trends and Outlook Survey, Updated March 28, 2024, https://www.census.gov/hfp/btos/data.
– Felten, E., M. Raj and R. Seamans (2021), “Occupational, industry, and geographic exposure to artificial intelligence: A novel dataset and its potential uses”, Strategic Management Journal , 42(12), p. 2195-2217, https://doi.org/10.1002/smj.3286.
– Filippucci, F., P. Gal, A. Leandro, C. Jona-Lasinio and G. Nicoletti (2024), “The impact of Artificial Intelligence on productivity, distribution and growth: Key mechanisms, initial evidence and policy challenges”, OECD Artificial Intelligence Papers, No. 15. OECD Publishing, Paris, https://doi.org/10.1787/8d900037-en.
– Georgieff, A. (2024), “Artificial intelligence and wage inequality”, OECD Artificial Intelligence Papers, No. 13, OECD Publishing, Paris, https://doi.org/10.1787/bf98a45c-en.
– Jones, C. (2023), “The A.I. Dilemma: Growth versus Existential Risk”, NBER Working Paper, No. 31837, National Bureau of Economic Research, https://doi.org/10.3386/w31837.
– Nordhaus, W. (2021), “Are We Approaching an Economic Singularity? Information Technology and the Future of Economic Growth”, American Economic Journal: Macroeconomics, Vol. 13/1, p. 299–332, https://doi.org/10.1257/mac.20170105.
– OECD (2018), Personalised Pricing in the Digital Era. https://www.oecd.org/competition/personalised-pricing-in-the-digital-era.htm
– OECD (2024), “Labour Market Shortages, Mismatches and Megatrends”, Global Forum on Productivity, forthcoming.
– Van Noorden, R. and J. Perkel (2023), “AI and science: what 1,600 researchers think”, Nature, vol. 621/7980, p. 672-675, https://www.nature.com/articles/d41586-023-02980-0.
– WEF (2023), “Jobs of Tomorrow: Large Language Models and Jobs”, https://www.weforum.org/publications/jobs-of-tomorrow-large-language-models-and-jobs/.