Realistic view on the economic impact of Generative AI
Introduction#
Growth rates in developed economies like the US have not increased despite the startling rate of technical advancement in recent decades. Many hurried to call the increased usage of digital services a turning point during the pandemic. However, as we stated at the time (and ever since), the predicted growth implications were not likely to occur, and they never did.
It is easier to see technology's future possibilities when one is aware of its past setbacks. Technology is just a fuel, which is one explanation for the modest performance. Effective technological adoption also requires a spark to spur productivity increase.
Lack of technology that can fully replace labor has been a barrier to better productivity growth, especially in labor-intensive services. With services reliant on asymmetric, mutual human connection, automation in manufacturing has no technical equivalent.
This is something that generative artificial intelligence (AI) hopes to change. However, in order to assess its potential influence with any degree of realism, we need to examine the mechanisms that link technology to increased productivity in general.
Pay attention to prices and costs, not apps.#
Productivity increase is all too frequently presented as the result of technology advancements leading to new product innovation. Significant productivity gains stem more from at-scale cost reduction than from new or improved goods, while both are crucial. The macroeconomic power of technology lies in its deflationary character.
We have drawn attention to the misplaced emphasis on slick apps at the expense of the real world of expenses by using the tale of the lowly taxi. Uber, Lyft, and Grab might be the driving forces behind society's advancement, both practically and figuratively, but where is the increase in productivity? Improving the ratio of inputs to outputs is the key to increasing productivity, as any first-year economics student understands.
Apps haven't completely changed that; labor and capital inputs—in this case, the driver and the car—remain the same, and driver-rider matching is somewhat improved. However, rising prices indicate that productivity hasn't changed; if it had, prices would have decreased.
Why? Businesses that can use technology to replace labor will cut prices to overtake competitors with greater costs in the market. Strong productivity growth occurs in the macroeconomy when that process moves through sectors.
The major change in transportation will occur if and when drivers are replaced by algorithms and sensors. Because you secretly settle payment or because the app shows the driver's location, it won't arrive.
The lack of tech's ability to ignite the technology-cost-price effect has been the reason for its lackluster growth impact. Impact is now more likely due to generative AI's ability to replace non-linear interactions in the service sector, such as contact centers, marketing, research, and design.
Businesses should be aware that the biggest winners from generative AI are customers.#
Customers will benefit if technology leads to strong productivity increase through cost reduction and declining prices. Real incomes will increase as prices decline, freeing up funds for other uses.
Think about how much money people used to spend mostly on food, but when costs dropped due to mechanization and fertilizers, eventually, people had more money to spend on other household items and services, like travel. This is how technology propels overall growth, proving that gloomy forecasts of widespread unemployment are unfounded as increased expenditure also generates employment.
This implies that for businesses, the tech-cost-price-income productivity cascade poses both a risk and an opportunity. Businesses who can maintain a competitive edge, cut costs, take market share, and lead the cost curve will succeed at the expense of others who cannot.
While generative AI will produce new corporate titans or revitalize those that already exist, some industries may believe AI jeopardizes industry revenues for all businesses.
This occurs when labor-saving technology is so easily available to all businesses. Reduced earnings and a pricing battle follow. Rather of resulting in soaring industry-wide profits, productivity advances in the auto sector, shipping, or aviation have instead produced cheap pricing, intense competition, and limited profits.
Therefore, the strategic implications of generative AI and other innovation for businesses are both offensive (cutting costs to acquire an advantage) and defensive (cutting expenses to remain viable).
Remain realistic about the macroeconomic effects of generative AI.#
A crucial component of a technological puzzle that also includes sensors, 5G, robots, biotechnology, and other elements, generative AI has the potential to increase productivity, but to what extent? Prodigious inventions are usually accompanied with euphoria, and some recent projections suggest that US productivity growth may increase by almost 300 basis points (bps).
That is overly flamboyant. Even though it would be tempting to extrapolate and generate macroeconomic predictions from bottom-up case studies, these estimations are still based on assumptions. Unpredictable obstacles that will extend timescales and restrict impact include public acceptability and regulatory friction.
The US economy would more than double its generally acknowledged trend growth rate, from about 2% to 5%, according to the estimate of 300bps. The same line of reasoning was used to forecast that increased digital use during the pandemic would increase productivity growth by 100bps or more, a prediction that proved to be incorrect.
Previous increases in productivity offer hints about likely effects. Is it possible that the current technological wave will resemble, surpass, or be smaller than the information and communication technology (ICT) boom that increased productivity in the middle of the 1990s and early 2000s? The last time availability and excitement for a new wave of technology coincided with a tight labor market to accelerate GDP by roughly 100bps for about ten years was then.
Conclusion#
This is something that generative artificial intelligence (AI) hopes to change. However, in order to assess its potential influence with any degree of realism, we need to examine the mechanisms that link technology to increased productivity in general.