In management, as in other disciplines, new technologies sometimes force us to reconsider the conclusions of established theoretical models. Frédéric Fréry writes that this was the case for digital platforms, which transaction cost theory had shown were intrinsically less efficient than integrated companies with their own assets and employees. Other advances, such as generative AI, may have the same kind of disruptive impact.
Research is sometimes ahead of practice. As far back as the early 1990s – some 15 years before Airbnb (2008) and Uber (2009) arrived on the scene – some scientific studies had already theorised about the emergence of digital platforms linking independent service providers with end customers. However, this hypothesis came up against a practical impossibility: transaction cost theory had shown that these digital platforms were intrinsically less efficient than integrated companies with their own assets and employees. Digital platforms were therefore a potential curiosity, but certainly not an organisational reality.
The unavoidable theory of transaction costs
The transaction cost theory stems from Ronald Coase’s insights, which were largely developed by the work of Oliver Williamson, for which both won the Nobel Prize in Economics (in 1991 and 2009 respectively). This theory provides an answer to this fundamental question: “Why are there companies rather than a fragmented market?” It justifies the existence of integrated companies on the grounds that markets involve specific costs, known as transaction costs: each time you want to do something on a market, you have to find financial backers, hire equipment and draw up contracts with service providers. The main advantage of companies is that these transaction costs only have to be paid once: financial backers, equipment and service providers are turned into long-term shareholders, investments and employees, respectively. As a result, companies are much more efficient when it comes to recurring activities.
There is another dimension to this. According to the work of George Akerlof, who was also awarded the Nobel Prize in Economics in 2001, one-off transactions between individuals who do not know each other engender mutual distrust that encourages cheating by both parties.
Overall, digital platforms have traditionally been reduced to nothing more than a theoretical curiosity: not only would no one agree, a priori, to host strangers in their home or to be driven around by a non-certified driver, but these approaches were also structurally less efficient than using integrated companies. Platforms therefore seemed to have no future. But all that changed profoundly in the 2000s, as a result of a series of technological developments.
Technology turns theory on its head
Thanks to the sea container and the internet, followed by the development of the web, smartphones and finally social media, it has become very inexpensive to coordinate independent service providers, find individual customers and set up digital platforms that make it easy to connect the former with the latter. In fact, the same arguments that led us to favour integrated companies can now be reversed. It was this reversal, which moved from a purely theoretical hypothesis in the 1990s to a practical reality a decade or so later, that legitimised the emergence of digital platforms such as Airbnb and Uber.
At the same time, we must pay tribute to eBay, which came up with the idea of asking its users, both buyers and sellers, to rate each other by awarding stars. This simple idea, scaled up massively by the spread of the web and then smartphones, and subsequently adopted by almost all digital platforms, circumvented Akerlof’s conclusion about the opportunism of transactions between strangers. Without smartphones and mutual evaluation, trust between strangers could never have been established and these business models could never have existed.
By drastically reducing transaction costs and providing a way to establish trust between strangers, a series of technological developments has led to a rebuttal of the conclusions of three Economics Nobel Prize winners.
How generative AI impacts other theories
Other management theories could see their conclusions refuted by technological developments, in particular generative artificial intelligences (AI) providers such as ChatGPT, Bard, Bloom and Copilot. There are three possible consequences:
One. This technology could turn the mythical Homo economicus into a reality where humans are capable of making decisions to maximise their utility in any situation. This would mean challenging much of behavioural finance or economics, while giving new operational effectiveness to conventional models. Generative AI is likely to give every investor, every customer and every manager the ability to consistently make optimal decisions in real life. If this is the case, it could invert a whole area of contemporary research.
Two. By the same token, because of how they work (generating the most likely text in a given context from a pre-established corpus), generative AI could be a tenfold reflection of bounded rationality, and therefore provide a much better understanding of the limits of human rationality. This could boost interest in finance and behavioural economics, but also in the study of consumer behaviour in marketing or individual and collective decision-making in management.
Three. We might even consider that generative AI is a good analogy for the way human thought works. Some psychological and sociological studies, along with research into cognitive biases, suggest that human thinking often follows an automatic, preconditioned path. Like ChatGPT, humans have a natural tendency to complete sentences with the words that seem most likely to fit the context. Generative AI could therefore provide a better understanding of how human intelligence works, which would undoubtedly have an impact on a number of established research outcomes in management.
For the moment, there is no way of predicting the impact of these technologies on management models, but it is likely to be significant. This re-reading of old theories in the light of new tools reminds us that models are only valid in relation to a certain state of knowledge. As the biologist Jean Rostand mischievously put it: “Theories come and go, but the frog remains.”
- This blog post is based on Digital generative AI platforms: when technology disrupts management models, part of ESCP Business School’s ‘new techs and the future of individuals, organisations, and society’ impact paper series..
- The post represents the views of its authors, not the position of LSE Business Review or the London School of Economics.
- Featured image provided by Shutterstock
- When you leave a comment, you’re agreeing to our Comment Policy.