Digital Economy Dispatch #179 -- Will AI Revolutionize the Economy? Here’s Why Experts Disagree

Digital Economy Dispatch #179 -- Will AI Revolutionize the Economy? Here’s Why Experts Disagree
14th April 2024

Erik Brynjolfsson is on a mission. He wants to prove to the world that AI is not “just another digital technology” but heralds a significant shift in productivity that will reshape business and society. And what’s more, he is willing to bet his reputation on it (and $400).

AI is rapidly transforming our view of the world and raising expectations about its effects on everything from employment and public service delivery to healthcare and entertainment. Yet, its impact on the economy is a topic of much debate. Some experts, like Erik Brynjolfsson, believe that AI is on the verge of causing a productivity boom. Others, like Robert Gordon, are more skeptical. Exploring both sides of this debate provides significant insight into the role AI may play as organizations seek to deploy AI-at-Scale and as we all try to make sense of AI’s disruptive force in our lives.

To demonstrate commitment to their points of view, Erik Brynjolfsson and Robert Gordon, both well regarded economists, have registered a $400 bet regarding the future economic impact of AI. Brynjolfsson believes that AI could drive significant productivity growth and economic gains in the coming years. In contrast, Robert Gordon believes that economic growth is stagnating and may not see the same kind of transformative impact from AI that Brynjolfsson expects. Part of a long standing debate between the 2 economists, they have made a $400 "long bet" on this disagreement about the future economic effects of AI.

What is this debate all about? And why do they disagree on whether AI will drive a major productivity boom and economic growth, or if growth will remain stagnant despite the rise of AI technologies?

Brynjolfsson's Optimism

Erik Brynjolfsson is an economist and director of Stanford University's Digital Economy Lab. He is a strong proponent of AI's potential to revolutionize the economy. Brynjolfsson argues that AI is fundamentally different from previous technological advancements. Unlike electricity or the combustion engine, AI has the ability to learn and improve over time. This, he believes, will lead to a sustained increase in productivity growth.

Despite the US Congressional Budget Office's prediction of slow productivity growth, Brynjolfsson argues that we are on the verge of a productivity boom. This optimism stems from three factors. Firstly, AI is revolutionizing various industries and accelerating discoveries. Secondly, general-purpose technologies like AI often take time to show their impact in official statistics because initial investments focus on intangible assets. This creates a "J-curve" where productivity dips before rising sharply. We are nearing the upswing of this curve for AI, he believes. Finally, temporary factors like pandemic relief and increased digital adoption will further boost productivity.

More broadly, Brynjolfsson claims we are on the cusp of a widespread AI-driven productivity boom. He points to the fact that AI is already being applied in a wide range of industries, from manufacturing to healthcare. These applications are automating tasks, improving decision-making, and creating new products and services. He also notes that the development of AI is accelerating. As AI algorithms become more sophisticated and powerful, their impact on the economy will become even more pronounced.

Gordon's Skepticism

Robert Gordon, an economist at Northwestern University, is more cautious about the potential of AI. In a widely cited book analyzing the US economy, he has argued that past periods of rapid productivity growth have been followed by periods of stagnation. For example, the productivity gains from electricity and the combustion engine were largely exhausted by the 1970s. Gordon worries that AI may not be as transformative as these earlier innovations.

Furthermore, Gordon highlights that the current productivity growth of 1.8% is part of a historical pattern of ups and downs. Recent figures, for example, show minimal impact from large investments in digital technologies. While the computer revolution brought a temporary surge, in terms of productivity data its impact seems to have faded by 2005.

Hence, Gordon doubts that robots and AI will be the next big productivity booster. For instance, a significant increase in robots hasn't revived manufacturing productivity, and AI hasn't revolutionized many sectors as many had hoped. He argues that many economic activities are so complex and embedded in organizational structures and processes that they simply cannot be radically transformed easily by AI.

Gordon also points out that AI has not yet led to a revolution in many sectors of the economy. For example, he believes that customer service jobs have been largely unaffected by AI. Additionally, many tasks simply cannot be automated, as they require human skills such as creativity and social intelligence.

The Macroeconomic Impact of AI

Both sides of this debate have merit. To understand their perspectives and put them into context, we must step back and consider the broader macroeconomic issues that AI raises.

One of the most useful perspectives comes from a 2023 study by the International Monetary Fund (IMF) explores the potential macroeconomic impacts of AI. The study, including participation from Erik Brynjolfsson, identifies three key areas of interest: productivity growth, income inequality, and industrial concentration. These three define the extent of the uncertainty that must be faced in predicting AI’s impact and underscore the wide variety of views on how AI’s future will unfold:

• The impact of AI on productivity growth is uncertain. The IMF study outlines two possible scenarios. In a low-productivity growth scenario, AI adoption is slow and limited to large firms. This could lead to further stagnation and missed opportunities. In a high-productivity growth scenario, AI is widely adopted and complements human workers, enabling them to focus on more creative tasks.

• AI could also exacerbate income inequality. If AI primarily replaces human labour, it could lead to a widening gap between the rich and the poor. However, AI could also create a more equitable future by assisting less-experienced workers and improving the overall work experience.

• The impact of AI on industrial concentration is also uncertain. If open-source AI models become available, smaller firms can leverage this technology and compete with larger ones. Conversely, if AI development remains expensive, it could further empower a handful of industry giants.

The Uncertainty Principle

The debate between Brynjolfsson and Gordon highlights a key question for digital leaders today: how will AI impact their organizations and the broader economy? The truth is that it is just too early to tell. There remains major uncertainty in key aspects of AI development, deployment, and adoption.

Given this doubt, we can extract three important lessons to inform today’s AI strategies:

1. Embrace Uncertainty: The true economic impact of AI remains to be seen. Leaders should be prepared for a range of possibilities, from stagnant growth to a productivity boom. Planning for both eventualities allows for agility and course correction as the landscape unfolds.

2. Focus on AI-Human Collaboration: Regardless of its ultimate impact on jobs, AI is most likely to augment human capabilities. Invest in strategies that leverage AI to empower your workforce, not replace it.

3. Prepare for Disruption: AI has the potential to disrupt entire industries. Leaders should continuously scan the horizon for emerging AI applications and be prepared to adapt their business models accordingly.

By understanding the ongoing debate and its potential consequences, digital leaders can develop robust AI strategies that position their organizations for success in the face of an uncertain future.