부산시청 도서요약
   글로벌 트렌드내서재담기 

책표지






    • Technology AI, Digitalization, & the Outlook for the U.S. Economy

    By Global Trends Editor Group 

    Favorable trends in U.S. demography are poised to complement the deployment of maturing AI and digitalization technologies.

    This combination will transform traditional industries and create new ones.

    For those who are ready, the resulting surge in productivity and ROI will represent the greatest wealth-building opportunity in human history.

    Understanding these trends can give managers, investors, and policy makers a huge advantage over those who fixate on short-term factors like interest rates, unemployment, and the current business cycle.

    In the August 2023 Trends issue, we explained why demographic realities portend enormous investment opportunities emerging through at least the mid-2030s.

    Given the demographic factors which are already locked in, we touched on the emergence of potentially transformative AI and digitization trends.

    In this issue, we want to further explore the role technology will play in shaping this extraordinary era.

    Understanding the interplay of technology and demography is crucial because they are both predictable, but they act in very different ways.

    Technology is ever-advancing and ever-accelerating.

    On the other hand, demography is cyclical, always progressing from birth to death to renewal.

    As we know, demographic cycles have existed throughout human history, and the implications of generational cycles have been researched back at least 500 years.

    Yet despite lots of human activity, standards of living improved at a glacial pace, hardly rising from century-to-century over thousands of years.

    Then, roughly 250 years ago, advancing waves of technology suddenly began lifting human affluence.

    For example, since our country’s founding, U.S. GDP per capita in constant 2022 dollars has risen from under $2,000 to more than $77,000, This technological advance and the accompanying economic progress were not random.

    It was embodied in a series of techno-economic revolutions , which built upon each other.

    For managers, investors, and policy-makers, it’s important to recognize that these technological revolutions are far more complex and nonlinear than the generational demographic trends.

    Instead of being born, living, and dying, each revolutionary technology starts with a unique innovation event referred to as a “technological big bang”.

    The most recent big bangs were the invention of Intel’s 4004 microprocessor in 1971 and Henry Ford is first assembly line in 1908.

    Later, after transforming the economy, every revolutionary technology becomes ubiquitous and commoditized.

    For instance, we could not live our modern lives without electricity, mass production, and railroads.

    Yet, while each of these technologies was once looked at with the amazement now associated with AI & robotics, all of them are now “taken for granted.”

    Another big factor in forecasting technology’s impact on jobs, consumers, and geopolitics is recognizing that earth-shattering progress enabled by each revolutionary technology occurs within a complex ecosystem.

    Moreover, its ultimate impact is shaped by the free market interplay of ecosystem participants creating an avalanche of innovations and wealth.

    That process obviously depends on competition among an initial set of industry participants, but it is further enhanced by the rise of new entrants and the impact of direct substitutes.

    At the same time, the industry benefits as its suppliers offer better inputs, and its customers develop more advanced needs.

    And ideally, this evolutionary process is encouraged and refined by smart government incentives.

    Yet, despite all of this complexity, history shows that truly transformational technologies such as railroads, mass production, or digital computing adhere to a well-defined pattern which we refer to as a “Techno-Economic Revolution.”

    As shown in the printable Trends issue, digitalization has conformed to this pattern since the 1970s.

    And we remain confident that it will continue to do so in the coming years.

    In deciding how to respond to change, it helps to understand how today’s extraordinary affluence was enabled by the five sequential techno-economic revolutions.

    And it’s important to recognize how each revolution was built upon the prior revolutions by harnessing a new and transformational, general-purpose technology: 

    First, industrial factories with steam power;

    Second, railroads and steam ships;

    Third, mega-process technologies including steel, electricity & chemicals;

    Fourth, mass production of automobiles, oil, and more;

    and Fifth, info-tech & digital computing.

    Crucially, each revolution follows a reliable pattern.

    It consists of an initial stage called installation and a subsequent stage called deployment.

    In the installation phase , the commercial value is derived primarily from the technology itself.

    This phase inevitably ends in a speculative frenzy during which the technology fails to deliver on excessive short-term expectations.

    This end is typified by “the dot-com crash” and “the crash of 1929.”

    Then, in the deployment phase, the full commercial value is realized primarily through using the technology to optimize nearly every aspect of the broader economy.

    Importantly, between the booms associated with the Installation and Deployment phases is a Transition phase which involves a malaise or depression.

    The most recent malaise occurred between 2000 and 2016, starting with the Dot-Com Crash, and including the Great Financial Crisis.

    Prior to that, the most recent malaise era was the Great Depression, which ended in World War II.

    Finally, the Deployment phase ends when the enabling general-purpose technology achieves full “Maturity.”

    This gives way to a period of weakness and confusion during which the next Techno-Economic Revolution germinates.

    The last period of prolonged weakness ran from around 1973 to 1982, when digital technology began to germinate and mass production technology had already reached saturation.

    Even though we’ve been in the Deployment Phase of the Digital Techno-Economic Revolution for roughly six years, its primary economic impact is just now being felt.

    Why is that? The full flowering of a techno-economic revolution requires many pieces working together to form a complex eco-system before take-off.

    Though no two techno-economic revolutions are identical, it’s useful to consider parallels between today’s situation and early parts of the Mass Production revolution’s Deployment phase.

    The Covid pandemic and social tumult of the current era have a lot in common with the late 1940s.

    The parallels include the enormous government debt, relatively healthy corporate and household balance sheets, surging inflation, serious labor shortages, housing shortages, international tensions, and high government interference in the private sector.

    Consequently, it was not until the early 1950s that the economy returned to a sustainable upward path resembling that of the 1920s.

    By analogy, the overarching economic challenge of the 2020s is to get back to a sustainable trajectory resembling the 1990s.

    Armed with that knowledge, it’s possible to understand the implications of transformative digital technologies like generative and analytical AI as well as robotics.

    This understanding helps us to anticipate threats and opportunities and prepare to address them.

    The key factor connecting technology and affluence is productivity.

    Output per person is defined by the number of hours worked and the output per hour.

    Without genuine technological innovation, you have only three options: 1) work more hours per person, 2) add more machines per person, or 3) find more clever ways to work.

    Unfortunately, each of these mechanisms for increasing wealth is quickly exhausted.

    However, breakthrough technologies permit us to deploy new machines that can produce more per hour of work, and it can unleash a whole new set of “ways to work more cleverly.”

    That is the secret that has permitted us to grow real per capita GDP by an astounding factor of at least 36-times in the last 225 years.

    That’s several times more than the cumulative progress over the prior five thousand years!

    At its root, each techno-economic revolution is based on one general-purpose technology such as the microprocessor, the assembly line, or the locomotive.

    However, it spawns a whole series of productivity-enhancing innovations during the installation and deployment phases.

    During the information technology revolution, those include the PC, World Wide Web, AI and robotics.

    In the 2020s, we’re approaching an AI-driven productivity inflection-point which resembles the take-offs we experienced with the widespread introduction of the PC in the 1980s and the World Wide Web in the 1990s.

    However, those “installation phase inflections” resulted primarily from the new technology letting people do things better, faster and cheaper.

    On the other hand, AI and robotics offer the opportunity to transform existing industries by doing things that previously couldn’t be done or that weren’t worth doing by traditional means.

    These include things like automated discovery of game-changing drugs and materials, self-driving cars & trucks, and job-site construction robots multiplying the capabilities of human craftsmen.

    What does that opportunity look like?

    A combination of analytic and generative artificial intelligence is poised to broadly super-charge productivity growth across the economy and around the world.

    As explained in the July 2023 Trends issue, McKinsey & Company estimates that AI could contribute up to $25.6 trillion a year to global GDP when fully deployed.

    This enormous opportunity comes from two varieties of AI: Analytic and Generative.

    The estimated pay-off of analytic AI is up to $17.7 trillion in added annual global GDP.

    The global payoff from generative AI is estimated at $7.9 trillion a year.

    And it will be made possible by rapidly escalating software capabilities, computing power, storage capacity, data collection and network bandwidth, as well as new business practices.

    Current and future economic realities depend on many factors.

    As we’ve explained, demography and technology are the most powerful.

    However, it also depends on behavioral trends which influence how business, consumers and government respond to the threats and opportunities created by demography and technology.

    We’ll discuss this aspect later in this issue.

    Here , it is more important to examine the potentially decisive role that technology will play in determining whether the “soft-landing scenario” highlighted in our sister-publication, Business Briefings, comes to fruition.

    A recent report from JPMorgan examined prior soft-landings and focused on the most recent one engineered by Alan Greenspan in the mid-90s.

    According to that report, “a tech-led investment spending boom, accompanied a sharp acceleration in productivity growth, fueled economic growth in the second half of the 1990s.”

    The report goes on to say, “the 1990s U.S. experience - in which a supply-side boost to both productivity and labor supply fueled growth - would seem to provide the best recipe for a soft-landing outcome.

    Indeed, the mid-1990s saw further margin expansion in the aftermath of Fed tightening alongside firming wages, strong growth and stable inflation.”

    James Pethokoukis of the American Enterprise Institute argues that, “AI-related tech spending could be the supply-side growth impulse that helps the Powell Fed manage to achieve a soft landing today.”

    As he observes, “there are some encouraging signs, including investors pouring money into generative AI startups in hopes of profiting from what could turn into a trillion-dollar sector over the next decade.”

    As JP Morgan notes in a separate report, it goes beyond that, “The investment prerequisite for the IT boom of the ‘90s was business investment in information processing equipment. In any prospective AI boom the investment prerequisite will likely be investment in software. And so far, the prospects in this cycle look promising.”

    For reference, the 90s productivity upshift was truly impressive, surging to an average of 3.1 percent from 1995 through 2004, after averaging just 1.5 percent over the previous quarter-century.

    History indicates that a decade-long productivity up-shift associated with AI is highly likely; the question is when.

    Yet despite this evidence, the current consensus view of U.S. economists is that AI will not boost productivity growth.

    In fact, they assume productivity will continue to slow on its long-term trend path.

    They attribute that to demographic trends, which we refute in the August Trends issue.

    And they also cite the growth of the counter-productive public sector addressed in trend #3 this month.

    Consequently, an AI-driven productivity boom would offer enormous upside that is underappreciated, especially by institutional investors, just as it was in the 1990s.

    The biggest argument for the consensus view is that it often takes longer than expected for businesses to adopt and efficiently use technologies to a sufficient extent for the gains to show up in the labor productivity data.

    An exhibit in the printable issue shows this timing issue as related to applications of devices driven by electric motors in the 1920s, as well as the household and workplace applications of PCs, eighty years later.

    In assessing the “development date” for this technology you should remember that even before ChatGPT, you could find pockets of optimism about the impact of machine learning on productivity and economic growth.

    Two years ago, Stanford University economist Erik Brynjolfsson began arguing that businesses had invested sufficient time and money in figuring out how to use AI to enable real productivity benefits from their initial investments.

    However, Brynjolfsson observes that new technologies like PCs and AI are typically subject to a J-curve effect, where there can be a productivity slowdown as investments are made, then there’s surge as those investments start to really pay off.

    And key economic evidence ignored by the consensus suggests that the productivity turning-point for AI is at hand.

    What’s the bottom line? The United States is poised to benefit from a convergence of demography and technology resembling the one we saw in the decades following World War II.

    However, few investors, managers or policymakers yet appreciate the magnitude of the emerging opportunities and threats.

    That gives those who correctly assess the situation a genuine competitive advantage.

    Given this trend, we offer the following forecasts, for your consideration.

    First, AI will add roughly $16 trillion to global GDP by 2030, disproportionately benefiting OECD countries.

    As with the McKinsey study highlighted in our August 2023 issue, Price-Waterhouse Coopers analyzed 300 use-cases by country and industry sector.

    Based on pre-Covid data, it concluded that China would benefit enormously as AI enabled automation of manufacturing.

    However, deglobalization and tech sanctions have totally flipped-the-script.

    In fact, AI and robotics will serve as a major enabler of reshoring, near-shoring and friend-shoring.

    North America, South Korea, Japan, and Europe will be the big winners as AI helps advanced economies deal with their growing skills shortages.

    Second, AI will create more jobs than it will destroy, especially in North America.

    Unlike most of the post-war period, the 2020s and 2030s will see demand for skills grow far faster than supply.

    The winners will be businesses and workers which are flexible and innovative.

    A recent Goldman Sachs study shows that the impact will be widespread, but only in a few cases will it eliminate all or most of an employee’s job.

    The reality is that certain tasks will disappear allowing one person to do the work of several.

    This will enable companies to deliver today’s value with roughly 25% fewer people.

    However, a rapidly growing economy with a shrinking labor pool will quickly absorb displaced people in new jobs.

    For employees, the key is to build the skills which enable them to move into better jobs rather than having to settle for downgrades.

    Unlike the era of hyper-globalization, the coming era will maximize growth in better-paying American industries.

    Third, accelerating rates of technology adoption will make the productivity surge associated with AI much faster than those associated with PCs or electric motors cited earlier.

    For instance, it took PCs over three times as long as tablets to achieve 50% penetration.

    Meanwhile washing machines, a major productivity enhancer enabled by electric motors, took 9 times as long.

    We expect the full adoption of game-changing AI solutions like autonomous automobiles, to be paced by systemic challenges rather than technology readiness.

    Fourth, adoption of generative AI will get a disproportionate share of attention because of its potential to simultaneously disrupt so many jobs and industries.

    According to research firms IDC and GroupM, these systems could realistically disrupt $100 billion in cloud spending, $500 billion in digital advertising and $5.4 trillion in e-commerce sales.

    Perhaps more than any other company, Google has reason to both love and hate chatbots.

    According to the New York Times, it recently has declared a “code red” because of Generative AI’s potential to undermine its $162 billion business of showing ads in searches.

    On the other hand, Google’s cloud computing business could be a big winner.

    At this point, mega-cap “tech companies” including Google, Microsoft and Amazon are in a race to provide businesses with the software and substantial computing power behind A.I. chatbots.

    The cloud computing providers have gone “all in” on A.I. over the last few months.

    They now realize that in a few years, most tech spending will be on A.I., so it is important for them to make big bets.

    And, Fifth, despite the hoopla surrounding the disruptive impact of generative A.I, roughly 75% of the payoff will come from analytic A.I.

    While applications in cyber-security and robotics are especially attention-getting, the real game-changers involve areas where human capabilities have increasingly picked all of the low-hanging fruit.

    Genomics, drug discovery, medical diagnostics, nanotechnology, brain science, materials engineering, chemical synthesis, and chip design are all moving beyond the point at which the unaided human mind can make rapid progress.

    Over the next two decades, entirely new technologies and industries made possible by AI-enabled science will improve our quality of life more than it’s been improved over the past century.

    Resource List
    1. Faster, Please! September 15, 2023. JAMES PETHOKOUKIS. How Will AI Help Us, Exactly?

    2. National Bureau of Economic Research. April 2023. Erik Brynjolfsson, Danielle Li, and Lindsey R. Raymond. Generative AI at Work. https://www.nber.org/papers/w31161

    3. The Economist. September 13, 2023. Economist. How scientists are using artificial intelligence.

    4. Harvard Business Review. June 06, 2023. Kathy Baxter and Yoav Schlesinger. Managing the Risks of Generative AI. https://hbr.org/2023/06/managing-the-risks-of-generative-ai

    5. CNN. March 29, 2023. Michelle Toh. 300 million jobs could be affected by latest wave of AI, says Goldman Sachs. https://www.cnn.com/2023/03/29/tech/chatgpt-ai-automation-jobs-impact-intlhnk/index.html

    6. Stanford HAI. 2023. Jack Clark, Ray Perrault, et. al. AI-Index-Report.

    7. Faster, Please! September 15, 2023. JAMES PETHOKOUKIS. Can AI prevent a US recession?

    8. MIT Technology Review. June 10, 2021. Erik Brynjolfsson & Georgios Petropoulos. The coming productivity boom: AI and other digital technologies have been surprisingly slow to improve economic growth. But that could be about to change.

    9. The New York Times. August 29, 2023. Steve Lohr. The A.I. Revolution Is Coming. But Not as Fast as Some People Think. https://www.nytimes.com/2023/08/29/technology/ai-revolution-time.html

    10. Goldman Sachs. March 27, 2023. Joseph Briggs & Davesh Konani. The Potentially Large Effects of Artificial Intelligence on Economic Growth.

    11. McKinsey & Company. June 7, 2023. Michael Chui and Lareina Yee. Generative AI could increase corporate profits by $4.4 trillion a year, according to new. research

    12. The New York Times. June 13, 2023. Andrew Ross Sorkin, Ravi Mattu, Bernhard Warner, Sarah Kessler, Michael J. de la Merced, Lauren Hirsch and Ephrat Livni. Accenture Makes a $3 Billion Bet on A.I. 

    13. The New York Times. March 8, 2023. Tripp Mickle, Cade Metz and Nico Grant. The Chatbots Are Here, and the Internet Industry Is in a Tizzy.

    14. Bloomberg Intelligence. June 01, 2023. Bloomberg Intelligence. Generative AI to Become a $1.3 Trillion Market by 2032, Research Finds.

    15. Pitchbook. August 24, 2023. Marina Temkin. Investors bet on generative AI to revolutionize drug discovery. 

    16. Bloomberg. September 1, 2023. Priya Anand. Paging Dr. GPT? Tech Investors Bet AI Finally Poised to Transform Health Care.

    17. THE WALL STREET JOURNAL. Sept. 4, 2023. Sam Schechner and Deepa Seetharaman. How Worried Should We Be About AI’s Threat to Humanity? Even Tech Leaders Can’t Agree.

    18. Trends. August 8, 2023. Trends Editors. Cashing in on America’s Demographic Windfall.

    19. Trends. July 4, 2023. Trends Editors. An AI-driven Productivity Boom is Coming.

    20. Trends. May 4, 2023. Trends Editors. Trends Users Unlock Their Full Potential.