By: Heather Kitchens
Edited by: Caled Al-Adsani
Graphic by: Arsh Naseer
Introduction
The commercialization of artificial intelligence (AI) has catapulted a world connected by the cloud into what some are calling an “AI Revolution.”1 According to Bloomberg Intelligence, generative AI has the potential to become a $1.3 trillion market by 2032.2 While generative AI is in the nascent stages, some researchers are forecasting AI-driven productivity may significantly boost economic growth in the U.S. and beyond by 2030,3 while others predict a longer time horizon, and some suggest AI may not have an appreciable impact on productivity statistics.
Diffusion of AI: A Historical Perspective
AI has been quietly permeating society over the last eight decades4 with recent worldwide attention on AI stemming from the practical applications of generative AI. AI first appeared in the 1950s and was formally recognized as a field of study at the Dartmouth Summer Research Project on Artificial Intelligence in 1956.5 In 2022, the public’s interest in AI rose exponentially with the commercial availability of generative AI tools.6 Generative AI has a variety of applications as “a type of artificial intelligence that can learn from and mimic large amounts of data to create content such as text, images, music, videos, code, and more, based on inputs or prompts.”7 In late 2022, the generative AI chatbot app ChatGPT emerged with over one million users in its first week.8 ChatGPT now has over 180 million users9 and has been classified as the “fastest-growing consumer application in history.”10 Reflecting the evolution of generative AI, Figure 1 shows that AI system test scores exceeded the human performance benchmark in reading comprehension, image recognition, language understanding, handwriting recognition, and speech recognition at various points in recent history.11 Researchers have noted that AI system performance outside of test environments varies.12
Figure 1: Test scores of AI systems on various capabilities relative to human performance13
Source: Kiela et al. (2023) – with minor processing by Our World in Data14
Labor Productivity and Technology
Labor productivity, defined as output per hour,15 has stagnated in the U.S. from 2006-2023, with the exception of 2009 and 2010. As depicted by the U.S. Bureau of Labor Statistics in Figure 2, the U.S. annual percent change in productivity in the nonfarm business sector from 1998-2005 exceeded the long-term average rate for 1947-2018 of 2.1%,17 with an average rate of 3.3%.16 However, from 2006-2018 the annual percent changes in productivity reflected low growth with rates primarily below 1.5% (Figure 2), which is lower than the long-term average rate for 1947-2018 of 2.1% and significantly lower than the average rate from 1998-2005 of 3.3%. Continuing this period of low growth, the U.S. average annual percent change in productivity in the nonfarm business sector from the fourth quarter of 2019 through the fourth quarter of 2023 was 1.6%18 (Figure 3), which is also lower than the long-term average rate of 2.1% from 1947-2018 and the average rate from 1998-2005 of 3.3%.19
Figure 2: Labor productivity growth: annual percent changes, nonfarm business sector, 1994-201820
Source: U.S. Bureau of Labor Statistics.21
Figure 3: Productivity change in the nonfarm business sector, 1947 Q1 – 2023 Q422
Source: U.S. Bureau of Labor Statistics.23
While a number of variables may be influencing the decades of below-average rates, some economists find that the slowdown in productivity despite years of innovation may be consistent with the “productivity paradox.” The productivity paradox is attributed to Nobel-prize-winning economist Robert Solow who stated in 1987 that “you can see the computer age everywhere but in the productivity statistics.”24 In response to the productivity paradox, economist Erik Brynjolfsson described the four explanations proposed for the paradox including output and input mismeasurement, lags caused by learning, redistribution where benefits are not observed at the aggregate level, and mismanagement of the IT.25 In addition, Brynjolfsson has argued that the impact of new technologies on growth relies on the invention of complementary investments that can result in “a productivity J-curve, where productivity initially falls, but then recovers as the gains from these intangible investments are harvested.”26 Some researchers find that the proposed explanations for the productivity paradox and the Productivity J-curve may explain the productivity slowdown.
In addition, the endogenous growth theory has been reported to show “that the short-run effect of the arrival of a new GPT can often be to reduce growth, by putting the economy through a long and costly adjustment period.”27 The potential for slower growth following the emergence of a new General Purpose Technology (GPT) may bolster the argument that generative AI, if determined to be a GPT, could impact the productivity statistics after a period of organizational and societal adjustments. GPTs are technologies that exhibit “pervasiveness, technological dynamism, and innovational complementarities”28 like the steam engine, electric motor, and semiconductor.29 GPTs are innovations in a technological canon for which some researchers believe generative AI should be included.
AI, Productivity, and GDP
Despite researchers’ concerns regarding whether the productivity paradox will persist, a number of experts anticipate that AI systems have the potential to drive a significant increase in productivity and, given “productivity growth is a significant driver of GDP growth,”30 an increase in gross domestic product (GDP). Price Waterhouse Coopers conducted an analysis to identify the potential “size of the economic prize”31 from AI, rather than a direct estimate of future economic growth. The analysis found that by 2030 AI could contribute up to $15.7 trillion to the global GDP, a 14% increase,32 which includes an estimated $6.6 trillion from increased productivity.33 According to Goldman Sachs Research, generative AI could increase the global GDP by nearly $7 trillion, a 7% increase, and increase productivity growth by 1.5 percentage points over ten years.34 Goldman Sachs analysts find that a significant economic impact from AI is possible but several years away stating that the “sizable increase in AI-related investment and large productivity gains among early adopters adds to our confidence that generative AI poses meaningful economic upside, while the slow adoption pace suggests that sizable macroeconomic impacts are still several years off.”35 In April 2024, the International Monetary Fund (IMF) World Economic Outlook reflected on the potential impact of AI on labor productivity, reporting that depending “on how widely it is adopted and whether it replaces or augments workers, the estimated global growth impact varies from 10 to 80 basis points in the medium term.”36 The IMF also reported that the impact of AI technologies on economic growth is “highly uncertain but potentially substantial”37 and found that “AI could prove transformative for medium-term global growth.”38
Potential AI Impact Across Sectors
Potentially bolstering the economic forecasts that have been described, researchers are finding that AI systems have the potential to drive efficiencies across various sectors. In the healthcare sector AI is assisting with early disease detection,39 in the transportation sector AI is forecasting passenger demand,40 and in the energy sector AI is forecasting energy usage and demand.41 In addition, in the manufacturing industry managers have recognized AI’s utility “in enhancing product and process innovation, reducing cycle time, wringing ever more efficiency from operations and assets, improving maintenance, and strengthening security, while reducing carbon emissions.”42 According to a McKinsey & Company study of 63 generative AI use cases, generative AI has the potential to add $2.6-$4.4 trillion in economic benefits annually when applied across industries.43 The McKinsey study found that generative AI has the potential to deliver value across industries through a variety of functions including generative AI-developed computer code, marketing and sales content, and customer service solutions.
Labor Productivity: Studies on Augmentation
In addition to the sectoral impacts that have been described, recent studies have evaluated the potential for AI systems to improve labor productivity. A study published by the National Bureau of Economic Research found that the number of customer issues resolved per hour by customer support agents using a generative AI-based conversational assistant tool increased by 14%, on average.44 A study by Harvard, MIT, and the Wharton School of 700 Boston Consulting Group consultants showed that highly skilled worker productivity can be improved by as much as 40% when using generative AI within the boundaries of its capabilities.45 Furthermore, an analysis by Accenture and Frontier Economics found that labor productivity could increase in the U.S. by up to 35% due to AI until 2035 (Figure 4). In addition to these studies, researchers are examining how AI systems will impact the labor workforce. Some researchers postulate that leveraging AI systems may not lead to massive job losses. However, other researchers emphasize the dichotomy between the democratization of AI benefits and the potential for job displacement.
Figure 4: Where AI is Aiding Productivity46
Sources: Statista, Accenture, Frontier Economics47
Investment in AI
While researchers are conflicted on whether AI will impact productivity, investment in AI has been steadily increasing from both a corporate investment and a global infrastructure perspective.
Corporate Investment: As shown in Figure 5, global corporate investment in AI has been rising steadily over the past decade. In 2013 the global corporate investment in AI was $14.57 billion and rose to $276 billion in 2021.48 Despite a decrease in corporate investment in 2022 to $189 billion,49 corporate interest in AI is continuing to grow with one survey showing 64% of businesses anticipate AI will increase productivity.50 According to Grand View Research, the global AI market is expected to have an annual growth rate of 36.6% from 2024-2030.51
Figure 5: Global Corporate Investment in AI52
Source: Chart: NetBase Quid 2022, Stanford University 2023 AI Index Report53
National Investment: In recent years, the global competitive landscape has been focused on national investments in AI infrastructure and developing a labor workforce with an AI skillset. According to the technology research and consulting firm Gartner, by 2027 “the productivity value of AI will be recognized as a primary economic indicator of national power.”54
A recent Brookings Institution analysis compared the readiness levels of various countries for achieving national AI objectives (Figure 6). The analysis identified the U.S. as the world leader in technology readiness from an AI infrastructure investment perspective,55 while rating the U.S. significantly lower on people readiness.56 The Brookings Institution analysis found India, Singapore, Germany, and China to be the highest rated in people readiness, as shown in Figure 6.57
Figure 6: Four Quadrants of National AI Strategy Implementation58
Source: Brookings Institution59
Additional Views on AI and Productivity
While some researchers forecast a nexus between AI and productivity, others believe AI will not have a significant impact on the productivity data. Robert J. Gordon, a Northwestern University economist, contends that technology must have a “profound effect on every aspect of human existence”60 to drive sustained productivity growth and does not believe AI meets that standard. Gordon expounded upon this theory stating, “The Third (digital) Industrial Revolution generated a productivity boost of only a decade between 1996 and 2006, as contrasted to the five-decade (1920-70) interval of rapid productivity growth following the Second Industrial Revolution, because the earlier inventions had a more profound effect on every aspect of human existence.”61 Gordon has argued in the past that “AI is nothing new, and its applications thus far do not suggest a revolutionary boost to productivity or mass destruction of jobs.”62
Other researchers find that a significant impact from AI on productivity may be possible, but requires systemic organizational changes that could drive a longer time horizon. According to the 2023 IBM Global AI Adoption Index survey, one in five businesses surveyed responded that they were not yet planning to use AI due to “limited AI skills and expertise, too much data complexity, and ethical concerns.”63
Conclusion
The future of AI is a paradox because the future is here while AI systems are iterating its future to come. The utility of AI is its greatest asset with its potential for bolstering creativity, productivity, and efficiency. Researchers do not agree on the productivity dividends AI may produce; however, experts agree that transparency and safety standards must be in place given the unknown and exponential power of AI systems. Some researchers emphasize that given the potential for AI systems to be transformative, depending on the diffusion timeline, it may not take long to determine the potential of AI to undergird the productivity statistics.
The opinions expressed are those of the author and do not reflect the opinions of their employer.
Bibliography
[1] Cheishvili, Archil. 2021. “The AI Revolution Is Happening Now.” Forbes. https://www.forbes.com/sites/forbestechcouncil/2021/08/25/the-ai-revolution-is-happening-now/?sh=532ed5d928c8.
[2] Bloomberg. 2023. “Generative AI to Become a $1.3 Trillion Market by 2032, Research Finds.” Bloomberg. https://www.bloomberg.com/company/press/generative-ai-to-become-a-1-3-trillion-market-by-2032-research-finds/.
[3] PwC.com. n.d. “What’s the real value of AI for your business and how can you capitalize?” https://www.pwc.com/gx/en/issues/analytics/assets/pwc-ai-analysis-sizing-the-prize-report.pdf. Price Waterhouse Coopers.com. Accessed April 19, 2024.
[4] Roser, Max. 2022. “The brief history of artificial intelligence: the world has changed fast – what might be next?” Our World in Data.org. https://ourworldindata.org/brief-history-of-ai.
[5] Dartmouth University. n.d. “Artificial Intelligence Coined at Dartmouth.” Dartmouth University. https://home.dartmouth.edu/about/artificial-intelligence-ai-coined-dartmouth
[6] Harris, L. A. 2023. “Artificial Intelligence: Overview, Recent Advances, and Considerations for the 118th Congress.” Congressional Research Service. https://crsreports.congress.gov/product/pdf/R/R47644.
[7] Harvard University Information Technology. n.d. “Generative Artificial Intelligence (AI).” Harvard University. https://huit.harvard.edu/ai.
[8] Holroyd, Stephen. 2024. “From the world wide web to AI: 11 technology milestones that changed our lives.” World Economic Forum. https://www.weforum.org/agenda/2024/03/11-technology-milestones-ai-quantum-computing-vr/.
[9] Tong, Anna. 2023. “Exclusive: ChatGPT traffic slips again for third month in a row.” Reuters. https://www.reuters.com/technology/chatgpt-traffic-slips-again-third-month-row-2023-09-07/.
[10] Harris, L.A. 2023. “Generative Artificial Intelligence: Overview, Issues, and Questions for Congress.” Congressional Research Service. https://crsreports.congress.gov/product/pdf/IF/IF12426.
[11] Kiela et al. 2023. – with minor processing by Our World in Data. “Test scores of the AI relative to human performance” [dataset]. Kiela et al., “Dynabench: Rethinking Benchmarking in NLP” [original data]. (from: Kiela, D., Thrush, T., Ethayarajh, K., & Singh, A. (2023) ‘Plotting Progress in AI’, Contextual AI Blog. Available at: https://contextual.ai/blog/plotting-progress (Accessed: 02 April 2024)). https://ourworldindata.org/brief-history-of-ai. Accessed April 19, 2024.
[12] Roser, Max. 2022. “The brief history of artificial intelligence: the world has changed fast – what might be next?” Our World in Data.org. https://ourworldindata.org/brief-history-of-ai. Accessed April 19, 2024.
[13] Kiela et al. 2023. – with minor processing by Our World in Data. “Test scores of the AI relative to human performance” [dataset]. Kiela et al., “Dynabench: Rethinking Benchmarking in NLP” [original data]. (from: Kiela, D., Thrush, T., Ethayarajh, K., & Singh, A. (2023) ‘Plotting Progress in AI’, Contextual AI Blog. Available at: https://contextual.ai/blog/plotting-progress (Accessed: 02 April 2024)). https://ourworldindata.org/brief-history-of-ai. Accessed April 19, 2024.
[14] Ibid.
[15] United States Bureau of Labor Statistics. 2021. “Overview of BLS Productivity Statistics.” https://www.bls.gov/bls/productivity.htm.
[16] Sprague, Shawn. 2021. “The U.S. productivity slowdown: an economy-wide and industry-level analysis.” U.S. Bureau of Labor Statistics. https://www.bls.gov/opub/mlr/2021/article/the-us-productivity-slowdown-the-economy-wide-and-industry-level-analysis.htm.
[17] Ibid.
[18] U.S. Bureau of Labor Statistics. n.d. https://www.bls.gov/productivity/home.htm. Accessed April 19, 2024.
[19] Ibid.
[20] Sprague, Shawn. 2021. “The U.S. productivity slowdown: an economy-wide and industry-level analysis.” U.S. Bureau of Labor Statistics. https://www.bls.gov/opub/mlr/2021/article/the-us-productivity-slowdown-the-economy-wide-and-industry-level-analysis.htm.
[21] Ibid.
[22] U.S. Bureau of Labor Statistics. n.d. https://www.bls.gov/productivity/home.htm. Accessed April 19, 2024.
[23] Ibid.
[24] Lee, Geunjoo, and James L. Perry. 2002. “Are Computers Boosting Productivity? A Test of the Paradox in State Governments.” Journal of Public Administration Research and Theory: J-PART 12 (1): 77–102. https://research-ebsco-com.proxy.library.cornell.edu/linkprocessor/plink?id=6d3ac129-3e89-34ba-8045-cac58ccb72d5.
[25] Brynjolfsson, Erik, Benzell, Seth, and Rock, Daniel. 1993. “The Productivity Paradox of Information Technology: Review and Assessment.” Communications of the ACM. 36. 66-77. 10.1145/163298.163309.
[26] Brynjolfsson, Erik, Benzell, Seth, and Rock, Daniel. 2020. “Research Brief: Understanding and Addressing the Modern Productivity Paradox.” MIT Work of the Future Task Force. https://workofthefuture-taskforce.mit.edu/wp-content/uploads/2020/11/2020-Research-Brief-Brynjolfsson-Benzell-Rock.pdf.
[27] Howitt, Peter. 2004. “Endogenous Growth, Productivity and Economic Policy: A Progress Report.” International Productivity Monitor, 8, 3-15. https://www.brown.edu/Departments/Economics/Faculty/Peter_Howitt/publication/IPM.pdf.
[28] Hogendorn, Christiaan and Frischmann, Brett M. (2020). Infrastructure and general purpose technologies: a technology flow framework. European Journal of Law and Economics, 50(3), 469–488.
[29] Ibid.
[30] Goldstein, Jeff. 2021. “Impacts from the pandemic on U.S. Productivity Growth.” The Atlantic Council. https://www.atlanticcouncil.org/commentary/blog-post/impacts-of-the-pandemic-on-us-productivity-growth/.
[31] PwC.com. n.d. “Sizing the prize What’s the real value of AI for your business and how can you capitalise?” https://www.pwc.com/gx/en/issues/data-and-analytics/publications/artificial-intelligence-study.html. Accessed April 19, 2024.
[32] Ibid.
[33] PwC.com. n.d. “Sizing the prize What’s the real value of AI for your business and how can you capitalise?” https://www.pwc.com/gx/en/issues/data-and-analytics/publications/artificial-intelligence-study.html. Accessed April 19, 2024.
[34] Goldman Sachs Research. 2023. “Generative AI could raise global GDP by 7%.” Goldman Sachs. https://www.goldmansachs.com/intelligence/pages/generative-ai-could-raise-global-gdp-by-7-percent.html
[35] Boughedda, Sam. 2024. “Goldman says AI revolution ongoing, but major impact still to be felt.” Investing.com. https://ca.investing.com/news/stock-market-news/goldman-says-ai-revolution-ongoing-but-major-impact-still-to-be-felt-432SI-3331565.
[36] International Monetary Fund. 2024. “World Economic Outlook—Steady but Slow: Resilience amid Divergence.” https://www.imf.org/en/Publications/WEO/Issues/2024/04/16/world-economic-outlook-april-2024.
[37] Ibid.
[38] Ibid.
[39] Marr, Bernard. 2023. “15 Amazing Real-World Applications Of AI Everyone Should Know About.” Forbes. https://www.forbes.com/sites/bernardmarr/2023/05/10/15-amazing-real-world-applications-of-ai-everyone-should-know-about/?sh=266e29485e8d.
[40] Ibid.
[41] Ibid.
[42] MIT Technology Review Insights. 2024. “Taking AI to the next level in manufacturing reducing data, talent, and organizational barriers to achieve scale.” MIT Technology Review Insights. https://www.technologyreview.com/2024/04/09/1090880/taking-ai-to-the-next-level-in-manufacturing/.
[43] McKinsey & Company. 2023. “The economic potential of generative AI: the next frontier.” McKinsey & Company. https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/the-economic-potential-of-generative-ai-the-next-productivity-frontier#industry-impacts.
[44] Brynjolfsson, Erik. Li, Danielle, and Raymond R., Lindsey. 2023. “Generative AI at Work.” National Bureau of Economic Research. https://www.nber.org/papers/w31161.
[45] Somers, Meredith. 2023. “How generative AI can boost highly skilled workers’ productivity.” MIT Management Sloan School. https://mitsloan.mit.edu/ideas-made-to-matter/how-generative-ai-can-boost-highly-skilled-workers-productivity.
[46] Ibid.
[47] Ibid.
[48] Lynch, Shana. 2023. “2023 state of AI in 14 charts.” Stanford University. https://hai.stanford.edu/news/2023-state-ai-14-charts
[49] Ibid.
[50] Haan, Katherine and Watts, Rob. 2023. “24 Top AI Statistics And Trends In 2024.” Forbes. https://www.forbes.com/advisor/business/ai-statistics/#sources_section.
[51] Grand View Research. (n.d.) “Artificial Intelligence Market Size, Share & Trends Analysis Report By Solution, By Technology (Deep Learning, Machine Learning, NLP, Machine Vision, Generative AI), By Function, By End-use, By Region, And Segment Forecasts, 2024 – 2030.” Grandviewresearch.com. https://www.grandviewresearch.com/industry-analysis/artificial-intelligence-ai-market. Accessed May 5, 2024.
[52] Lynch, Shana. 2023. “2023 state of AI in 14 charts.” Stanford University. https://hai.stanford.edu/news/2023-state-ai-14-chart.
[53] Ibid.
[54] Afshar, Vala. 2023. “The future of generative AI: Here’s what technology analysts are saying.” Zdnet.com. https://www.zdnet.com/article/the-future-of-generative-ai-heres-what-technology-analysts-are-saying/
[55] Dawson, Gregory S. and Desouza, Kevin C. 2022. “How the U.S. can dominate in the race to national AI supremacy.” Brookings Institution. https://www.brookings.edu/articles/how-the-u-s-can-dominate-in-the-race-to-national-ai-supremacy/.
[56] Ibid.
[57] Ibid.
[58] Dawson, Gregory S. and Desouza, Kevin C. 2022. “How the U.S. can dominate in the race to national AI supremacy.” Brookings Institution. https://www.brookings.edu/articles/how-the-u-s-can-dominate-in-the-race-to-national-ai-supremacy/.
[59] Ibid.
[60] Gordon, Robert J. 2018. “Why has economic growth slowed when innovation appears to be accelerating?.” National Bureau of Economic Research. https://www.nber.org/system/files/working_papers/w24554/w24554.pdf.
[61] Ibid.
[62] Ibid.
[63] Thomas, Rob. 2024. “AI is pervasive. Here’s when we’ll see its real economic benefits materialize.” Fortune. https://fortune.com/2024/04/01/ai-artificial-intelligence-when-real-economic-benefits-materialize-tech/.
Author Biography:
Heather is a Brooks Public Policy Fellow at the Cornell Jeb E. Brooks School of Public Policy and is pursuing an Executive Master of Public Administration at Cornell University. She has extensive experience in government contracting and holds an MBA with a focus on international business. Her research interests include public-private partnerships, government management, and digital governance.