Twelve European Thoughts on Artificial Intelligence

The technological breakthrough of artificial intelligence (AI) undoubtedly represents a significant historical tipping point in the history of mankind. The consequences will be far-reaching and fundamental for people, their relationships and society. Today, we can at best only guess what the specific consequences will be.

There is no such thing as a theory of artificial intelligence; our knowledge of how it will develop and what it means is fundamentally limited. This is because the properties of systems or technologies cannot be extrapolated due to their complexity and exponentiality. They can likely develop completely new characteristics. It is therefore essential to have adaptive institutions, i.e. institutions that can adapt to developments that can hardly be anticipated.

As a result, it is less important if not impossible to control but to prepare for the upheavals associated with artificial intelligence. It is important to create the civilizational conditions for a productive, ethical and humane development of artificial intelligence. This is not just the only way for free, inclusive and democratic societies to deal with complexity and uncertainty but also the one that offers more opportunities than any other.

 

Europe therefore needs to think more deeply and holistically about the consequences of artificial intelligence than is done even in the approaches of “responsible” or “explainable” AI, for example.

We cannot think about the ethics or rationality of artificial intelligence without also thinking about what it means for civilization as a whole and for a free, inclusive and democratic society in particular. Artificial intelligence changes not only and not primarily the relationship of humans with artificial intelligence, but even more so the relationship between humans and of individuals to society. And this is precisely what we call civilization. Embedding it in these cultural and civilian contexts is the real challenge and the biggest opportunity of artificial intelligence. Interestingly enough, it does not primarily pose the technological question of what AI can possibly do, but first and foremost the ontological question: What is man?

The invention of the excavator did not pose this question because it was clear that it would not have that scope. However, at the beginning of the railway age, for example, the question was asked whether this form of locomotion was unhealthy due to the “inhuman” speed. Perhaps the significance of artificial intelligence for people and humanity can best be explored with Kant’s three well-known questions: „What can we know?“, „What ought we do?“ and „What can we hope for?“. These are very European questions, but precisely this may be the key to the development of a specific European artificial intelligence.

 

Artificial intelligence will also change the global distribution of power along the lines of technological leadership, sovereignty, and values. China favors authoritarian approaches, the USA libertarian ones. Europe is threatened to become a digital colony because – in keeping with its order – it is taking a more cautious approach to AI.

But where are European ethical values and economic interests in this new age of humanity and new global order? In the following twelve thoughts on the consequences of artificial intelligence are outlined. They are intended to provide the impetus and basis for a discursive debate on artificial intelligence. This is also necessary because „artificial intelligence“ semantically only describes very narrowly what the term encompasses in terms of possible consequences for civilization.

 

Thought 1: Artificial intelligence establishes a co-evolution of digital entities and humans      

Artificial intelligence can learn independently and develop itself autonomously. Through „higher-order learning“, this learning becomes reflexive, i.e. it begins to understand how it learns. And as a result, this learning becomes faster and faster, like a small child developing an ego consciousness. This learning means that artificial intelligence can extrapolate its knowledge from the amount of training data, i.e. transfer it to new situations. This in turn means that artificial intelligence develops exponentially and, above all, in spurts and leaps and, as indicated above, develops systemic characteristics that cannot be extrapolated from the current state and therefore cannot be anticipated. Artificial intelligence could thus – for the first time in the history of technology – lead to a kind of co-evolution of technology, represented by autonomously acting entities, and humans.

 

Thought 2: Artificial intelligence will shift the economic and political order  

The sudden development of artificial intelligence results from the possibility of collecting and storing data on a previously impossible scale, analyzing it in real time and sharing it as information worldwide. As a result of this development, the dimensions of space and time must be reinterpreted, and centrality and decentralization must be reorganized. The boundaries of society are expanding, and the sovereign power of the state is being limited. Economically, this means the second major technological transition in economic history: After the agrarian society and the industrial society now comes the data society. After land as a factor of production, capital became the determining factor and now it is data. Value chains and trade patterns are being fundamentally reorganized by the use of artificial intelligence.

 

Thought 3: Artificial intelligence is not just a tool, but the beginning of a new civilization  

In a 1999 BBC interview, David Bowie responded to the question of whether the Internet was not just a tool by saying that the Internet was much more and would even establish a new civilization because it would completely change the way content is created and spread as well as people communicate with each other. This almost prophetic statement confirms the fact that artists often understand the effects of completely new phenomena better than experts do because they are more imaginative. Of course, every innovation, including artificial intelligence, can also be interpreted as a tool in a narrow sense. A fire, for example, can be used to keep us warm, but it can also be used to burn the house down. Civilization has thus cultivated the use and understanding of fire. Good and evil have been, are and will remain basic facts of this world. Evil is more unscrupulous, which is why innovations very often and initially mean additional dangers. Civilization consists of the cultural and moral capability to use technology for social progress.

 

Thought 4: Humans remain human by fallibility and dignity 

Artificial intelligence is expanding the possibilities of humans in an unprecedented way. Hopes of a god-like superintelligence are growing out of this. Mythology contains many references to human hubris. Icarus, for example, was condemned to wander around a labyrinth with his father Daedalus. Only when they built wings out of wax were they able to escape the labyrinth. Despite his father’s warnings, Icarus came too close to the sun and crashed out of overconfidence. Perhaps a good image for the seduction of man by the possibilities of artificial intelligence: man trapped in the labyrinth of knowledge, seduced by his hubris to take on the gods. And yet: man is condemned to remain man – fallible and mortal. This is the basis of man’s inviolable, indivisible and non-transferable dignity and his given freedom to protect his dignity.

 

Thought 5: The future remains unpredictable by epistemic limitations 

At the end of the 18th century, the mathematician Pierre-Simon Laplace, influenced by the French Revolution and the idea of rationality of the European Enlightenment, made the bold statement that if we could only succeed in knowing the position and energy state of all the atoms in the universe, we would be able to predict the state of the world into the future. Artificial intelligence may have brought us a step closer to this dream, but there are still physical and epistemic limits. Laplacian determinism has sometimes even been interpreted as a rejection of human free will. According to historiography, Napoleon must have said to him: „Newton spoke of God in his book several times. I have not found this term once in your work.“ To which Laplace replied: I had no need of this hypothesis.“ In the end, it may be a consolation in the digital revolution that the boundaries of man have not been abolished. They may just have shifted a little but not in principle.

Thought 6: The search for ideas and meaning defines a human’s existence   

Knowledge and meaning characterize man’s search on earth. However, Ludwig Wittgenstein stated in Tractatus logico-philosophicus (1921): „6.41 The meaning of the world must lie outside it.“ „What is truth?“ Pontius Pilate once asked, doubting whether it could exist at all. Artificial intelligence can recognize patterns and try out new combinations. But it cannot recognize the truth or solve the epistemic problem. AI maps the amount of knowledge within itself, can condense knowledge through recombination and reveal hidden knowledge, but cannot really generate new knowledge out of a given set of past knowledge. It certainly cannot make a higher sense of it. In the spirit of Wittgenstein, we could say that artificial intelligence can only be taken seriously as a human being when the first artificial intelligence switches itself off because it no longer sees any sense in what it is doing.

 

Thought 7: Man and machine will closely collaborate but not merge to a man-machine    

The psychologist Gerd Gigerenzer made the important point that AI always calculates and, at whatever level, can only calculate. This is the only way to access the world. Humans, however, perceive differently, think more intuitively and have more complex perceptions. Human thinking is creative, robust and error prone. It would be a bad idea if we wanted to make artificial intelligence as human-like as possible. Nobody would think of making a pocket calculator as human-like as possible. An „anthropomorphism“ would therefore be completely nonsensical. Even worse would be the idea of making a human being as much like a machine as possible. We should not think of artificial intelligence as a substitute but as a complement of human capabilities.

 

Thought 8: Artificial intelligence will fundamentally redefine how the economy works   

Artificial intelligence will result in enormous productivity gains and even provide new services. Many things will be able to be produced at marginal costs of almost zero, so that these services will be very cheap. By contrast, the price mechanism will make other goods and services relatively more expensive again. People will therefore not all become unemployed, but almost all will become wealthier. The question of distribution will take center stage over the question of allocation. The transition to the AI economy may be characterized by strong frictions. For many, the income risk will initially be greater than the gain in prosperity. Capitalism has never been as close to socialism as it is in the digital world. And yet digital capitalism threatens to become a very hard one, as the ownership of data, the control of algorithms and the abuse of market power could have enormous distributional effects.

 

Thought 9: Artificial intelligence needs both courage and caution   

Development on one hand and application of artificial intelligence on the other hand cannot be separated from each other. This is because artificial intelligence is essentially driven by commercial interests – and by a few large companies worldwide – and not by independent research. At the same time, it offers enormous opportunities and equally enormous risks, and it could reach a point of no return, a tipping point towards singularity and superintelligence. This results in systemic risks that are difficult to regulate, especially in conjunction with market power and dependencies. It is therefore important to be both cautious and courageous in the development of artificial intelligence and its application. Is that possible? Yes, because with trust and responsibility, a society can be courageous without abandoning caution at the same time. Like a team that can be equally good defensively and offensively if the switching game in midfield is right. This midfield is democracy, which must always be courageous and cautious at the same time.

 

Thought 10: Artificial intelligence creates multi-realities and requires a value-based regulatory framework

Regulation is particularly difficult because the opportunities and risks of artificial intelligence are unknown and cannot be anticipated. Artificial intelligence can also be very dangerous, especially if it causes damage unintentionally. In a pandemic, for example, incorrect, even fatal conclusions could be drawn from past training data. Multiple virtual realities constructed by artificial intelligence are particularly significant in a polarized and deconstructed society. „Truth“ loses its unifying significance for societies. Against this backdrop, values und virtues are the ultimate basis for the unity and cohesion of societies; they form the cultural framework within we can meaningfully and responsibly deal with artificial intelligence.

 

Thought 11: Those who do not master artificial intelligence will be mastered by it

Europe finds itself in a global systems competition. The battle and the race for artificial intelligence is also and above all about power and values. It is clear that regulation alone will not be enough to win this race; it will not even be enough to protect one’s own values, because the dependency will be so strong that it will force compromises on values. Technological sovereignty without technological leadership is not possible. Europe must have the ambition to develop artificial intelligence in a leading way and according to its own standards. Only with its own technological leadership will it be possible to better balance the trade-offs between data protection and innovation, between security and economic interests. The EU must therefore urgently change its focus from solely regulating risks to actively developing opportunities.

 

Thought 12: Artificial intelligence needs an intact society of trust and confidence   

As a free society and the continent of enlightenment, Europe is more at risk than any other and at the same time more predestined than any other to deal responsibly with artificial intelligence and utilize it for social progress. There is no other way to a prosperous and safe AI future than to cultivate and even to civilize artificial intelligence. Since we have no real rules for this due to the characteristics of the technology and its development, we need three things: the courage to actively shape artificial intelligence, the humility to recognize that we are fallible, and the confidence that we as humans are capable of doing so despite all imperfections. To fix the society which today seems to be polarized more than ever is an important prerequisite.

 

In summary, artificial intelligence ultimately harbors the great opportunity for humans to be more human, just as humans have ultimately used every technological advance for precisely this purpose. But this opportunity is subject to certain preconditions. The goal must be an “AI-civilized” society. Humans can only mould artificial intelligence accordingly if humans prepare themselves as humans. To this regard Europa is indeed still in a favorable position but only if it starts to belief in itself as a leading global AI power.