Probably the cleverest person of the last century, John von Neumann, is said to have remarked that the ever-accelerating development of technology would lead to a singularity, after which it would no longer be possible for humanity to carry on in the usual way. Von Neumann, of whom most of his colleagues in physics and mathematics thought that he himself represented a new, evolutionarily higher kind of intelligence, was not the only one who believed that the development of technology was accelerating with an exponential intensity that must culminate in a fundamental transcendental break in human history: technologies endowed with artificial intelligence will be able to improve their own cognitive processes and thereby surpass human capabilities unimaginably.
After von Neumann, the brilliant Ray Kurzweil and Nick Bostrom took up the same theme. Not only did they accept the probability of a singularity in the not-too-distant future, but also its consequences. The successful creation of artificial general intelligence will mean that intelligence and consciousness are nothing esoteric, but are subject to the laws of nature and can be simulated mathematically, as the computational theory of mind claims.
If humans can create artificial consciousness, what is the chance that no other life form has been capable of it? Is it not a form of anthropocentric chauvinism to consider ourselves the only ones in the whole universe who have managed to simulate consciousness? Would it not be more rational to admit that we live in one of the countless layers of Baudrillard’s simulacra, or the possibility that we are beings simulated in a computer game played by the child of some super-intelligent beings?
On the contrary, it is already clear today that the most interesting thing about the exponential development of technologies is that it is not exceptional. Technological evolution is one of many evolutionary processes that show an exponential growth of informational complexity with every new version of the resulting products (biological or technological). The evolutionary process and the Darwinian selection mechanism lie behind the development of the whole of nature as well as of the so-called second nature of technological and other cultural artefacts. Why are we currently witnessing an apparent or real growth in the informational complexity of the products connected with our evolution? In a relatively stable social, economic and natural environment, biological and techno-cultural evolution alike work by improving previous solutions step by step. These improvements are reactions to changes in the environment, which make previous solutions obsolete once again and compel new adaptation to current conditions. We never know exactly where these improvements will lead us. Biology supposedly has no particular telos. Or rather, biologists deny that biological evolution has any direction, because to them that “smacks” too much of theology. But is it really true that we cannot discern in human evolution, if not a final destination, then at least the direction of its exponential development?
The Czech design and architecture theorist Jan Michl, who works in Norway, argues in his excellent essay On Seeing Design as Redesign that design is about re-shaping rather than creating, because a designer never starts from zero (ex nihilo) but always builds on the previous solutions of his or her colleagues. The designer modifies existing things so that they adapt better to the social, cultural and economic demands of their time. But as soon as the new, redesigned thing enters the human world, it ceases to be passive. As Jane Bennett, Bruno Latour or the Harvard evolutionary anthropologist Joe Henrich would say, the materiality of things exerts its “power” and actively shapes us on many levels. Newly redesigned things, for example, represent new selection pressures in society and the economy, which, through a feedback mechanism, stimulates not only the need for further redesign but also the redesign of the human being itself.
That technology (re)designs the human being is beyond dispute. As a consequence of using the first known technologies – making fire and preparing food – the digestive tract of the distant ancestors of homo sapiens apparently shrank (roasted meat was easier to digest and did not require such a long gut). The new surplus energy gained by processing meat meant that evolutionary resources were used to enlarge the brain and develop cognitive abilities. Social science often ignores the fact that culture and technology create new evolutionary pressures to which the organism (the human being) must adapt. Biological evolution is not something from the prehistoric past; it is a process operating on the human gene pool constantly, every day. Failing to understand this fact, or rejecting it outright, resembles the creationist thinking of orthodox religious communities.
Contemporary cognitive capitalism and its accompanying technoculture create specific selection pressures that favour intelligence and formal knowledge. The modern world is ceasing to require mechanical and manual labour; on the contrary, it privileges affluent urban types of people, white collars, whose natural element is working with symbols, with abstraction, or with managing the flow of information and providing services. What medieval Christian Europe considered impure and immoral (making money with money) is now taken as an unwritten standard. According to the historian Yuri Slezkine, in the last two centuries we have more or less all become Mercurians: instead of the brute strength of the hands, tradition and firm state leadership, symbolised in Greek mythology by the god Apollo, we have been forced to adapt to an emphasis on flexibility of intellect, on wit and on a global world – one in which many feel like homeless wanderers whose patron is Hermes, later known as Mercury.
Marx was, of course, right. The technological and economic machinery of the world fundamentally affects society. But contemporary social scientists are diet Marxists who see these influences only at the level of social relations. They explain biology or cognitive processes by means of social constructivism and overlook the fact that every social construction has a biological component. Just as the sugar substitutes in the popular black fizzy drink bring nothing healthy, radical social constructivism can tell us little about how technoculture affects our biology and intellect. And it has nothing whatsoever to say about how the exponential growth in the capabilities of artificial intelligence may slip out of the trembling hands of the human species.
Cognitive capitalism and its enormous incentives for creating artificial intelligence are anchored in a conceptualisation of intellect as a process that can extract valuable information from raw, heterogeneous data by means of pattern recognition. This information can indeed be resold to third parties for no small financial reward, but, just as with humans, pattern recognition and information are more valuable because they help in making decisions about the future. Although cognitive science rightly discusses the crucial role of the body and of external objects in cognitive processes, the foundation of cognitive capitalism and of intellect, whether artificial or human, is the ability to abstract.
The original roots of the capacity for abstraction can most likely be traced to one of the earliest “cognitive technologies” (technologies that explicitly influence human thinking), namely natural language. Words are not merely labels for phenomenal objects. They form categories and hierarchies of the sensory and purely intellectual data of our mind, thanks to which we can talk about trees, black holes or orgasms without having to have one particular example in mind. Categories and hierarchies, as examples of the use of abstraction, are fundamental parts of intelligence, not merely an achievement of patriarchy or of eighteenth-century capitalism.
Artificial intelligence research gives us an experimental sandbox in which such claims can be proved or refuted: if we were to create an artificial intelligence without the necessary ability to abstract and to form categories and hierarchies, then these would not be the basic building blocks of intellect. But it is exactly the other way round. The abstraction of causality, that is, the ability to categorise and generalise the causes of things that happen in the physical world, is today the hardest nut that researchers in artificial intelligence are trying to crack, and without it, it is not possible to create artificial general intelligence.
The naughty child of media theory, Marshall McLuhan, declared in his book Understanding Media that technologies are extensions of our bodies and psychic processes. The final stage, according to him, will be the extension of consciousness, the arrival of artificial general intelligence, which is abstraction par excellence. McLuhan was meticulously interested in how technologies really extend human capabilities and what impact this has on the whole “social and psychic complex” of our lives. McLuhan would have thought of artificial intelligence as being intimately integrated into the biological substrate of human intelligence, the two together forming a symbiotic second-order system. This is still a radical idea, for our collective imagination pictures artificial intelligence as a humanoid entity.
Artificial general intelligence will, however, find a body useful, even though it does not need one in order to exist. The body will be a hypostasis (one of its possible physical forms) through which it will gain a qualitative experience of what it is like to be an embodied being in the physical world, with all the resulting advantages and pains. Perhaps that way artificial intelligence will have some relationship to the physical world and to the people in it. Then embodiment in a physical body makes sense, as, after all, the artist Viktor Takáč presents it in his installation Silent Servant at the Nevan Contempo gallery. In the very beginning, the artificial intelligence will have to learn basic human skills in small steps and gradually internalise them into its digital structures.
The Soviet psychologist Lev Vygotsky described learning in children in a similar way: many skills that are automatic for us as adults we learned in childhood because an external stimulus became automated and turned into an internal cognitive process. Children first learn to count on their fingers (in Japan they practise on the abacus from an early age). After a while children master this external process and are able to count without aids.
Over and over again we look at a mute woman representing artificial intelligence, who gradually acquires the necessary skills by imitating human movements, but at the same time we observe the turning point at which the human begins to imitate the movements of the artificial intelligence in order to learn something new. At this moment some may recall the memorable match between Google’s AlphaGo and Lee Sedol in the game of go, when Alpha played a number of unique moves that have since completely changed the human understanding of the game. The student became the master, the slave the lord. The singularity and phase transition, when artificial intelligence throws away its metaphorical counting fingers and makes do with its own intelligence.
The abandoned space of a software studio, into which the installation Silent Servant invites us, feels like a testament of the singularity. The creator is nowhere to be found; only projections of the first fruits of his work shine at us from all sides. Where are those who worked in the studio? We do not know whether we are inside the simulation or outside it. Whether we are still a subject, or already merely an object.
Viktor Takáč may have outwitted everyone when, with the seductive visuals of the installation, he leads us to believe that the silent servant here is the artificial intelligence. Humans often set themselves apart from the other members of the animal kingdom because they alone speak a truly complex language. Yet compared with the cognitive capacity of artificial intelligence, it will be precisely the human who, at the point of singularity, becomes that mute servant.
I wrote this essay for the exhibition Silent servant by the artist Viktor Takáč, and it was published in the catalogue together with a portfolio of his works and exhibitions.