The impressive capabilities and unprecedented growth of ChatGPT have released the hype all over again. Technology is about to disrupt white collar jobs, customer support, education, writing, and even mighty Google itself. Maybe. But before getting too excited, it’s worth noting that we’ve heard these predictions before only to see that, while individual companies rise and fall, dramatic industry shifts are relatively rare. Even the most affected sectors tend to carry on. Telephone companies, book publishers, news organizations, advertising agencies, record labels and retail stores are all still here. Airbnb didn’t destroy the hotel business, taxis still roam the streets, and banks haven’t been bypassed. Businesses and industries are not stupid; they can learn and adjust. Innovations tend to be more additive than zero-sum.

As amazing as ChatGPT is, its overall economic impact remains uncertain. Consider that the leap from paper documents to instantaneous Internet searches changed information access much more than the jump from Google to ChatGPT, and yet online search engines have had relatively little effect on the fundamental workings of health care, finance, education, government, and other information-intensive businesses.

AI’s potential to supplant Google is arguably a more likely scenario, but even in that case there is reason for caution. ChatGPT can produce remarkably well-written and clever answers to an extraordinary range of questions. But because its answers are sometimes wrong, strange, out of date, or bland, its use is problematic in areas where verifiable accuracy, explicit sources, current information, and differing opinions are important. From an information industry perspective, content providers are particularly concerned that ChatGPT often uses their information without acknowledgement, let alone links or compensation.^[1] Of course, Google, Microsoft, Baidu, and others are developing their own generative capabilities, and these will likely complement, rather than replace, their existing search engine offerings.

Looking ahead, search engine competition should increase substantially, weakening the case for antitrust intervention. Indeed, the long pattern of U.S. antitrust cases being launched at the very moment the market leader’s power has begun to recede risks being repeated.^[2] ChatGPT-type offerings will surely help people who don’t write well and increase the productivity of those who do. Their potential in creating content, developing software, translating languages, and other tasks is hard to overestimate. But just as Google search has greatly increased the capabilities of countless individuals without making them obsolete, generative AI is much more a powerful new tool than a radical economic and employment shift, at least for now.

Definitional Divides

We tend to overestimate the extent of industry and employment change because the term “disruption” has two very different meanings. A typical dictionary definition is “the action of preventing something, especially a system, process, or event, from continuing as usual or as expected.”^[3] In this everyday English-language sense, digital technology and the Internet have been highly disruptive as many industries, organizations, and individuals have had to significantly change the way they work. Using this definition, ChatGPT and its successors will also prove disruptive in many areas.

But within business schools, the technology industry, and many policymaking circles, the phrase “disruptive innovation” has a much more specific meaning. It is shorthand for a process of creative destruction where one or more new players deploy a radically different technology or business model to obsolete—or seriously diminish—the leaders in a particular market. This definition was famously articulated by the late Clayton Christensen in his seminal 1997 book, The Innovator’s Dilemma.^[4]

Christensen made the important distinction between disruptive versus sustaining innovation. They can be equally potent, but the former undermines established market leaders, while the latter supports them. Christensen-style disruptive change has occurred repeatedly within the IT industry, most famously in the 1980s and ‘90s when personal computers and local area networks replaced the once formidable minicomputer industry. Similar disruptions have also occurred outside of the IT sector, but less frequently—with bookstore chains, cameras, and CD/DVD stores among the most iconic examples. If ChatGPT effectively addresses its current shortcomings and takes down Google, that would be yet another example of disruption within the IT sector, but disrupting non-IT sectors remains a much higher and less likely bar.

So, depending on whether one uses the dictionary or Christensen definition, industry disruptions are either very common or relatively rare. These two entirely different usages explain why there has always been a lot of talk about industry disruption even though, in the Christensen sense, disruption has actually been in a lull in recent years, as blockchains, the Internet of things, self-driving cars, expert AI systems, et al., have yet to come close to supplanting existing market leaders. Few economic terms are used so loosely and confusingly, and much of today’s ChatGPT disruption talk reflects the fact that both meanings are so well entrenched.

The gap between these two definitions will be especially important over the coming years. If we use the dictionary meaning, Western societies will face massive disruptions during the 2020s—moving away from fossil fuels; electrifying transportation; developing more sustainable products and services; reducing dependencies on China; coping with climate change, current and potential military conflicts, and more. Virtually all of these challenges are taking place within the physical world, and thus digital technologies are likely to play a secondary role.

But whether these physical-world shifts will be disruptive in the Christensen sense remains to be seen. Thus far, incumbent market leaders have responded to changes in the automobile, energy, agricultural, and other markets. However, it’s still early in the game, and as Tesla and the Chinese EV-maker BYD have shown, leadership by new western firms or those from China, India, or elsewhere is entirely possible, with potentially massive consequences for global competition. One thing is clear: Given that both disruption definitions continue to be widely used, confusion about the extent of industry change won’t go away anytime soon. We should all use the term “disruption” more carefully, but we almost certainly won’t.

Endnotes