These days, hardly anyone is still talking about smart fridges which will order milk for you before it runs out. From a technical perspective, it is not a problem to fit a fridge with sensors that can read labels on food packaging. But the fact that we all still happily look in the fridge to see whether there is enough milk for breakfast shows that the smart fridge story, now more than 20 years old, seems to belong to the genre of the unfulfilled techno-fantasy.
Research into the history of technology has shown that most technological predictions and visions remain castles in the air. Most predictions are creatively reinterpreted to make them look in hindsight as if those investigating the latest trends were practically clairvoyants. A closer look, however, reveals that correct predictions of technological developments remain very rare exceptions. The histories of trend research, technological impact assessments, and future projections are ones of confusion and error.
This, however, does not make futurologists, visionaries, or forecasters any more humble – on the contrary, they are currently getting especially excited about artificial intelligence, digitalization, and robotics. There is talk of virtual shopping assistants that know the needs of their users better than the users themselves, of teams of robots that will soon win the football world cup, and of communication between internet chatbots without human beings even noticing it.
The use of Big Data, intelligent algorithms, and learning machines is predicted to bring about a fundamental change in the way work is organized in businesses, public administrations, hospitals, and universities (Mayer-Schönberger & Ramge 2018). It is expected that new forms of cooperation will emerge in which employees are no longer confined to the ghetto of their sections or divisions. Decision-making will no longer be the privilege of managers; rather, ‘robo-bosses’ will develop business strategies based on algorithms. Organizations as we know them will dissolve thanks to the technological possibility of providing products or services through a network of one-person enterprises, which are very cheap to coordinate.
The race for the most disruptive scenario
There is a competition to be the announcer of the most dramatic disruptor. You are considered a bore if you compare recent developments in artificial intelligence and Big Data merely to the invention of the first mainframe computers in the mid twentieth century. Rather, the significance of technological change in digitalization is said to equal that of the invention of the steam engine, the discovery of electricity, or the creation of the printing press. Alternatively, the progress made in artificial intelligence is compared to the Cambrian explosion in the fossil record 550 million years ago.
Of course, especially in artificial intelligence and robotics, even the visionaries have to admit that most of the predictions made so far have remained science fiction. But this has only led to an adaptation of the narrative. While they may admit that many of the predictions about artificial intelligence have not been borne out, they add that the enormous increase in computing power we are seeing now will soon change this. In a short while, the new story goes, researchers will solve the problems they have been trying in vain to crack for many years.
The problem with predictions is often not that the developments they describe are not technologically feasible. Although insiders joke that technology which works gets a ‘proper name’ while everything else is called ‘artificial intelligence’, in many cases it is not that the technology cannot be developed but that it cannot be successfully implemented.
Most technological revolutions fail due to cost-benefit analysis
At the purely technological level, currency systems could be fundamentally changed using the blockchain, but the energy costs would make individual transactions incredibly expensive, building trust in such currency systems would take a long time, and states would still attempt to intervene. The fridge that communicates with a delivery service does not fail to materialize because of technological problems but because the complexity of the cost–benefit analysis would drive most consumers back to conventional fridges.
The main problem with realizing technological visions is path dependence, the force which is so significant that new – and often superior – technologies cannot prevail. There is plenty of evidence that intelligent, decentralized power generation at the level of households would be more efficient than the existing centralized power generation. But because of the investment in the centralized network, there is an inertia that blocks change.
There are good reasons why fantasies run wild in the field of artificial intelligence. Futurologists and trend researchers earn their money by painting pictures of possible futures with as much detail as possible. Describing possible futures in the most radical terms allows consultants to persuade their frightened customers that they need more of their services. And for newspapers constantly searching for sensational headlines, colourful visions of the business of the future often make for better copy than descriptions of the actual changes taking place.
We should focus on the effects that already exist
In light of the understandable excitement of visionaries, forecasters, and trend researchers, we urgently need to inject some realism into the discussions about artificial intelligence and robotics. Instead of intense speculation about technological possibilities, upcoming products, or new organizational forms, we should analyse the effects of the new technologies that already exist.
In concrete terms, this kind of realism means that managers’ PowerPoint presentations would be rigorously purged of all elements of science fiction and instead filled with descriptions of the new technologies that are actually being used in their businesses. Consultants would no longer be determined to ask managers about trends that they can blow up into major threats for their next publication but would instead concentrate on the detail of the concrete technological changes taking place. When writing about ‘possible’ technological developments, journalists would make clear that their story belongs to the genre of science fiction.
It would be naïve to assume that there will be no more fundamental technological changes. In some sectors of the economy, digitalization is changing business models considerably. The technological interconnection between value-added processes, and also across organizational boundaries, is accelerating. Automation in production as well as in the service industry has progressed a great deal. The history of capitalism is always also the history of changes in the market driven by innovation. But it is precisely for this reason that it is so important to be realistic when describing new technologies.
Viktor Mayer-Schönberger and Thomas Ramge, Reinventing Capitalism in the Age of Big Data, New York: Basic Books, 2018.
