D.he digital technological upheaval in which we find ourselves naturally brings about extremes. These include expectations of salvation on the one hand as well as demanding apocalyptic on the other hand, which does not necessarily conjure up the impending doom of the world, but at least the end of humanity as we know it. Well advised is someone who does not adopt both, but instead reads Gerd Gigerenzer, for example.
The psychologist has published a book entitled “Click” that is didactically well done, is pleasantly factual and spans a wide range of artificial intelligence (AI), as it currently works and appears in well-known offers, about our handling of probabilities and Uncertainty through to social scenarios and personal recommendations.
How to develop a model of the world
Gigerenzer uses striking examples – such as dating platforms or automated driving – to open up a broad spectrum of what is currently known as AI. It illustrates which errors the corresponding algorithms can make – and which misinterpretations, in turn, people are prone to. This is not only instructive, it also suggests that a purely intuitive understanding of modern information technology is hardly sufficient for those who want to remain digitally confident in the future.
Gigerenzer then leads on to the question of whether it is possible to teach a computer what we call “common sense”: Even small children develop a model of the world by constantly observing and experiencing their surroundings; they subconsciously internalize the existence of natural laws, understand them what space and time are, what norms apply in those groups in which they are, and they can think causally, i.e. perceive what is cause and what is effect.
More computing power and data volume?
AI pioneers such as Turing Award winner Yann LeCun hope that this general understanding of things is essential that we can learn new things very quickly, that we only need a double-digit number of driving hours to get our driver’s license and not have to look at tens of thousands of pictures of lions in order to recognize this animal species in previously unknown images and from other perspectives.
This idea is not only inspiring the leading IT companies to construct much larger artificial neural networks and train them with even more data. However, experts are also certain that the solution does not just consist of more computing power and data volume. Gigerenzer illustrates this, among other things, by showing what it takes to really understand language.
A dedicated warning
Less convincing, however, are his statements about the conditions under which AI systems are superior to humans and vice versa. It seems to have been proven that computers can perform superhumanly in common games, i.e. clearly defined rule areas such as go or chess. It is quite arguable, however, that the human advantage, as Gigerenzer distinguishes it from it, automatically lies in dealing with uncertainty, with rule breaks, in which, as he remarks with reference to the economist Frank Knight, something like judgment, intuition and ” the courage to make decisions ”.
Aren’t the ability to make judgments and intuition also come from experiences, i.e. ultimately from data combinations? “If the future and the past are not the same, collecting and analyzing big data – which always comes from the past – can lead to wrong conclusions,” he continues. Probably true. But doesn’t all applicable human knowledge at any given point in time also come from the past personally or from other people and in some form recorded and passed on?
Gigerenzer’s book is much more than an introduction to statistics, AI and human technology behavior. He adds a staunch warning of the Internet’s legacy and its consequences. He complains about the business models of social networks and search engine providers based on targeted advertising and how these affect everyday life between parents and children or between road users in interaction with modern mobile phones. Based on the work of the economist Shoshana Zuboff, he also speaks of surveillance capitalism and sees it as a development for society as a whole that endangers individual freedom and health. He also describes the technical control that states like China and India are striving for. He sees this approach to the existential challenge of the western democracies in a global system competition.
Gigerenzer’s pessimistic view of the possible consequences of digital networking should be countered, for example, by access to knowledge and fellow human beings. Regardless of your point of view, the book is a successful all-rounder that will enrich the debate about a world whose importance continues to grow.
Gerd Gigerenzer: “Click”. How we stay in control in a digital world and make the right decisions. Translated from the English by Haine Kober. C. Bertelsmann Verlag, Munich 2021. 416 pp., Ill., Hardcover, € 24.
.