It's already begun! What is the "Great Divergence" of the AI Revolution?

Last time, we discussed opportunities for the advertising industry based on Christensen's "The Innovator's Dilemma." This second installment examines AI from the perspective of "Macroeconomics" – the broadest framework within this series structured as "Macro (Economy and Industry) ⇒ Micro (Service Image)." What exactly is the "Great Divergence"?

Industrial Revolution and Technology: AI as General Purpose Technology (GPT)

According to Professor Tomohiro Inoue of Komazawa University, industrial revolutions are driven by General Purpose Technologies (GPT). GPT refers to "technology that impacts all industries and triggers a chain reaction of complementary inventions." The GPT of the First Industrial Revolution (1770–1830) was the steam engine, while the GPTs of the Second Industrial Revolution (1865–1900) were the internal combustion engine and the electric motor. According to Kenneth Pomeranz, a "Great Divergence" (※5) occurred around 1800 between Western Europe, the United States, and Japan, which adopted GPT and shifted toward industrialization, and the rest of the world, which did not [Figure 1].

Manufacturing saw efficiency gains and productivity increases through GPT (First and Second Industrial Revolutions), but service industries, including intellectual labor, experienced only limited productivity growth. Since the 1990s, however, advances in computing have gradually begun to streamline service industries (e.g., bank tellers replaced by ATMs), and AI is expected to accelerate this trend (e.g., bank operators replaced by natural language dialogue systems). In other words, AI is significant because while steam engines and electric motors took over human "physical labor," AI takes over human "mental labor." "Standardized service industries" are the most easily streamlined by AI. It is well known that in the US, AI is already replacing paralegals (legal support roles). Note that the business efficiency gains from IT since the late 1900s are sometimes called the "Third Industrial Revolution," while the anticipated impact of AI on industry from now into the future is sometimes called the "Fourth Industrial Revolution." For example, since around 2011, the German government has championed "Industry 4.0," aiming to realize manufacturing in smart factories centered on CPS (Cyber-Physical System, to be explained in Part 6), a mechanism utilizing AI and IoT technology. At this year's World Economic Forum Annual Meeting (Davos Conference) in January, the theme was responding to the Fourth Industrial Revolution enabled by CPS [Figure 2]. (The author believes that since the First and Second Industrial Revolutions were revolutions in new energy, for the Third and Fourth to be "true" industrial revolutions, new energy and technology must be integrated.)

The Second Great Divergence: The Major Crossroads Already Underway

Mr. Inoue refers to an economy where productivity increases solely through automation and machine learning, without human intervention, as the "AK-type production economy." This is because its production function is represented as "Y=AK," where Y is output (GDP), A is the level of technology, and K is capital (such as investment in machinery). The production function for the economic growth model of advanced nations like Japan today, following the "Great Divergence," is called the "Cobb-Douglas model," expressed as "Y=AKαL(1-α)" (where L is human labor). While deep mathematical understanding isn't necessary here, the key point is that the AK-type production function driven by AI lacks the "L (labor)" present in the Cobb-Douglas model. In other words, "economic growth can occur without labor." In a Cobb-Douglas production economy, the economic growth rate is nearly equal to the rate of technological progress (g). Indeed, Western nations and Japan have maintained average annual growth rates of around 1-2%. Surprisingly, however, in an AK production economy, it has been pointed out that this very "growth rate figure" of 2% itself could grow exponentially. In other words, a world previously unimaginable—where growth averaging 2% annually could become 2% in year one, 4% in year two, and 8% in year three due to the AI revolution's autonomous machine learning and automated production—theoretically becomes possible. [Figure 3] (As will be detailed next time, economic growth cannot be realized without demand commensurate with technological progress. Therefore, the main argument here is that "the advertising industry should take on the role of creating demand itself").

Drawing from past industrial revolutions, Inoue calls this phenomenon the "Second Great Divergence," stating that "countries (organizations) that rapidly adopt AI as GPT will dominate the world economically, leaving others behind." (He notes that until around 2030, a hybrid of Cobb-Douglas and AK production models will persist.) Realistically, observing the rapid global growth of a limited number of players—Google, Facebook, Apple, Amazon, etc.—who are investing massive amounts in AI, one can already sense that the "Second Great Divergence" is underway. Salim Ismail of Singularity University calls such companies Exponential Organizations (※6). If the current advertising industry is a mature service sector, it will be most susceptible to AI's impact. Whether or not companies swiftly adopt AI technology will become a major fork in the road (great divergence) for future growth.

※5: Kenneth Pomeranz (translated by Minoru Kawakita), The Great Divergence: China, Europe, and the Making of the Modern World Economy (Nagoya University Press, 2015)

※6: Salim Ismail, Exponential Organizations (Diversion Books, 2014)