One of the most important issues for investors and policymakers alike is understanding the potential impact that generative artificial intelligence (Gen AI) models will have on the economy and labor market. McKinsey & Company uses the term to describe algorithms that can be used to create new content including audio, code, images, text, simulations, and videos. For lay people, they are often thought of as machine-learning techniques.
The task is daunting considering the complexity and sophistication of the models deployed, the wide range of applications they serve, and the inherent uncertainty about how they will evolve.
Chat GPT's Impact Likened to that of the Printing Press
Amid this, one thing is clear. Namely, the launch of Chat GPT one year ago has captured the attention of technology experts. Rob Thomas, IBM’s chief commercial officer, has called it a “Netscape Moment” referring to the introduction of the firm’s browser in 1994 that “brought the internet alive.” Kai-Fu Lee, CEO of 01.AI, likens it to the advent of Guttenberg’s printing press that made it possible for ideas to spread around the world at previously unimaginable speeds, which created huge gains for mankind.
The development of Gen AI has also resonated with equity investors. The so-called “Magnificent Seven” stocks, posted outsized gains last year, with those in the AI space leading the way. They include semiconductor giant Nvidia (NVDA) and Meta Platforms (META) that were up by nearly 240 percent and 195 percent, respectively.
Many Still Cautious of the AI Revolution
Some observers, however, caution that the AI revolution may not come as fast as optimists think or be as transformative as proponents contend.
Steve Lohr, who covers technology for the NY Times, observes that there has always been a lag between the invention of new technologies and their adoption across industries and the economy. He also cites the example of JPMorgan Chase, which has told workers not to put any bank information into the chatbot or other Gen AI tools. The reason: Management is cognizant of risks of leaking confidential data, and it questions how the data will be used and how accurate the answers will be.
Governments, moreover, will have an important say in determining regulatory policies for artificial intelligence. The European Union announced a landmark deal on AI regulation in December that the EU’s internal market commissioner proclaimed is: “much more than a rulebook—it’s the launch pad for EU startups and researchers to lead the global race for trustworthy AI.” According to The Economist, however, the launch of ChatGPT is making it more difficult for the U.S., Britain and the EU to agree on a common risk-based approach to regulate AI.
What is Gen AI's Potential to Boost Productivity?
One consideration that differentiates it from previous technologies is the much wider specter of jobs it can impact. The list includes professional workers in high-end, high-pay jobs, as well as people working in low wage, repetitive jobs, and in back offices.
Using a database that lists about 900 occupations, Goldman Sachs’ economists estimate that roughly two thirds of U.S. occupations are exposed to some degree of automation by AI. They further estimate that of those occupations that are exposed, roughly one quarter to one half of the workload could be replaced. All told, Goldman’s researchers conclude: “They could drive a 7% (or almost $7 trillion) increase in global GDP and lift productivity growth by 15 percentage points over a 10-year period.” If so, annual productivity growth for the U.S. would double from about 1.5 percent to 3.0 percent, which would be an unprecedented achievement.
Daron Acemoglu and Simon Johnson of MIT, however, are skeptical of this assessment. They see it at odds with the historical record in which new technologies that expand the set of tasks performed by machines and algorithms often wind up displacing workers and boosting corporate profits, but do not lead to shared prosperity.
An NBER working paper by David Autor and Anna Salomons helps shed light on this issue. Using four decades of cross-country and industry data, they find that automation typically displaces employment and reduces labor’s share of value added in the industries in which it originates. However, these job losses are often reversed by indirect gains in consumer-oriented industries and by increases in aggregate demand. Overall, they find that “technological progress is broadly employment-augmenting in the aggregate.” Indeed, the report observes that 60 percent of today’s workers are employed in occupations that did not exist in 1940.
A Brookings report by Martin Neal Bailey, Erik Brynjolfsson, and Anton Korinek spells out the case for an AI-powered productivity boom. The authors observe that large language models such as ChatGPT are powerful tools “that not only make workers more productive but also increase the rate of innovation, laying the foundation for a significant acceleration of economic growth.” The authors produce a simulation that shows that the difference in output between estimates based on future technology versus current technology results in nearly a doubling of output over 20 years.
At the same time, Erik Brynjolfsson, who directs Stanford’s Digital Economy Lab, acknowledges that the increase in productivity is unlikely to be straight-line. In fact, there is typically a period where productivity declines and there is a lull. The reason is that technological change often forces organizations to reinvent themselves and to develop new processes that take time.
Western & Southern Financial Group View
The view of the technology specialists at Western & Southern Financial Group is that generative AI has potential to drive meaningful productivity gains—both in speed/effectiveness for skilled knowledge workers and in processing efficiencies for more routine tasks. They are already seeing concrete examples in action with software development and creation of marketing materials.
Regarding the lag between adoption of a new technology and its impact on productivity, they observe that initial gains can be had very quickly without a high degree of technical proficiency, mainly due to the ease of chat-based interfaces such as ChatGPT and Bing that contain a broad and accessible body of knowledge.
Finally, our tech specialists see Gen AI as different from many emerging technologies that require deep expertise, significant investment and large scale integrations to drive gains. While these attributes will still bring disproportionate returns from investing in GenAI, they are not required to get into the game. This suggests that some type of broad incremental bump in productivity is possible from investing in it, although the extent will vary with individual businesses.
Past performance is not indicative of future results. This publication contains the current opinions of the author but not necessarily those of Fort Washington Investment Advisors, Inc. Such opinions are subject to change without notice. This publication has been distributed for informational purposes only and should not be considered as investment advice or a recommendation of any particular security, strategy, or investment product. Information and statistics contained herein have been obtained from sources believed to be reliable and are accurate to the best of our knowledge. No part of this publication may be reproduced in any form, or referred to in any other publication, without express written permission of Fort Washington Investment Advisors, Inc.