The Future of the Artificial Mind

Alessio Plebe, Pietro Perconti

• Publisher: CRC Press
• Year: 2022
• ISBN:
  9781000614701
• Edition: 1
• Topic: Computer Science
• Subtopic: Computer Science General
• Index: Computer Science

The Future of the Artificial Mind is about the social and technological challenges posed by the new wave of artificial intelligence, both from a technical and a cognitive perspective. Deep neural networks have brought about tremendous technological improvements. This renaissance in artificial intelligence, after decades of stagnation, has enabled new technologies capable of surpassing human performance, as in the case of visual recognition.

The book reviews the key ideas that have enabled these goals to be achieved and their historical origins. The book also considers some of the ethical and social challenges that the future development of artificial intelligence will face. Will humans fall in love with future android dolls? What will artificial sex be like? And what will it be like to travel in cars that will treat us as passengers instead of drivers? But predicting the future appears more magic than science. But when it comes to artificial intelligence, it is a constant temptation. Since it is well known that "the only way to get rid of a temptation is to enjoy it!", the hypothesis considered in the last chapter is that emerging trends point to a near future in which intelligence will be ubiquitous, but it will be difficult to identify its bearer. We may be heading towards an era of widespread intelligence, but an intelligence without accountability.

Chapter 1 The Landscape

1.1 Into Eega Beeva's Toolbox

If you are a millennial, you probably take for granted things that were futuristic to those born before the 1980s. You probably assume that when you consult a news website, it comes in real-time from all over the world. You take it for granted that, with a twenty-hour or so flight, any corner of the planet can be reached. You take for granted that, with your cell phone, you can talk to, see, and exchange messages with people anywhere. We take things for granted that were just science fiction scenarios after World War II. If you are not a millennial, as a kid, you may have fantasized about strange drawings where you could see flying cars, cities floating in the air with thousands of ethereal streets running through them, fantastic glasses that let you see the underwear under girls’ dresses, guns that could hit their target with laser beams capable of disintegrating any object in an instant. None of this fantastic world created for the imagination of boys after World War II was then realized as expected. But we knew that, after all. Nor, by the way, did we expect the cowboy world of the past to take the very form with which we entertained our imaginations at the time.

On the other hand, the future held for us just the sort of surprises we fantasized about as kids, though perhaps not quite in those areas. Global and digital communication, high-speed transportation, armaments, technological medicine are just a few of the achievements that have made many people’s lives almost futuristic. We live twice as long as our ancestors and, if we go to a hospital, we are investigated by machines that can literally see inside our bodies. After all, the dirty nerd glasses have been replaced by something more useful. There aren’t the laser guns we imagined, but the military launches attacks thousands of miles away via airplanes that fly without pilots and hit their targets from heights of thousands of feet without being seen by anyone. Cars don’t yet fly, except in a few prototype cases, but they now travel alone, gradually transforming our role from driver to passenger.

Figure 1.1: Eega Beeva by Bill Walsh and Floyd Gottfredson, from The Man of Tomorrow, 1947.

Few things have played a greater role in shaping the contemporary imagination than Walt Disney’s characters and stories. This is also true in the realm of imagining the future. In a 1947 issue of Mickey Mouse, a character appears who is destined to affect the most common way of imagining the future and the role that science would play in shaping it: Eega Beeva (proper name: Pittisborum Psercy Pystachi Pseter Psersimmon Plummer-Push). Pluto and Mickey Mouse are forced by bad weather to take refuge in a large cave where they meet a strange individual who seems to come from the future. In fact, the title of Mickey Mouse’s issue was just that: The Man of Tomorrow (Walsh and Gottfredson, 1947). Eega Beeva is not entangled in the quarrels of humans. He has a big head, a disarming confidence in the future and a bunch of extraordinary tools to pull out whenever needed. Taking a peek into Eega Beeva’s toolbox would be very useful to embark on the journey that this book proposes to its reader.

In this book, we would, in fact, like to investigate what form the mind takes as we go deeper into the future. We would like to show how cognitive science is leading us to conceptualize the mind as something profoundly different from what we expected. But, more importantly, we would like to understand how the mind is transforming into something artificial. The future of the artificial mind is, therefore, the subject matter of this book. As we will see at the end of the path, it seems that the distribution of the mind in the social environment and the loss of a subject that owns it are the main features of the transformation to which intelligence is subjected in the historical phase in which we are living. Perhaps we will have to live in a world in which intelligence will be everywhere, but there will be no one to own it. It is too early to say whether a world characterized by distributed intelligence, but with no one in charge, will be better than the current one. After all, these are only predictions and, as happened in the case of X-ray glasses and fMRI, things in the future will take an amazing shape, but not exactly in the sense we had predicted, or hoped.

If we are interested in the shape that human and artificial minds will take in the future, casting a glance into Eega Beeva’s toolbox would be of great help. Put another way, what is the right vocabulary to tackle the artificial intelligence journey? How should we use the words ‘mind’, ‘artificial’, ‘intelligence’, ‘representations’, or ‘computation’? It is important to note the modesty with which this last question is inspired. It has, in fact, no foundational intent. This book makes no foundational claim. It is not like Whitehead and Russell’s Principia Mathematica, in which the proof that 1 + 1 = 2 does not appear until page 379. The minimal convention on vocabulary just mentioned has the sole purpose of making explicit that a book devoted to artificial intelligence and cognitive science necessarily includes a number of prejudices regarding the use of numerous expressions and words.

1.2 Mind

The word ‘mind’ refers to a set of cognitive capacities such as attention, perception, consciousness, imagination, computation, judgment, and language. How each of these capacities develops and establishes relationships with the others constitutes the shape of a particular mind. Hence, the mind of a rabbit will be different from that of a human being, that of a child from that of an adult, that of a robot different from the collective intelligence of the internet, whatever that latter expression may mean. Having a mind, however, does not only mean having some cognitive capabilities, but also experiencing the exercise of those same capabilities. Having a mind means experiencing perspective in the first person, having a subjective point of view. On the other hand, having a mind is the condition for having not only a personal identity, but also a social one. The social roles that we play in our lives are the result of complex social attributions based on the use of intentional vocabulary, made of ‘beliefs’, ‘desires’, ‘emotions’, and ‘promises’.

The mental phenomenon is investigated from numerous scientific perspectives such as psychology, linguistics and artificial intelligence. Philosophy makes its contribution to the modern science of the mind primarily by attempting to answer the following three questions: What is it? Where is it? What is it for? The first question is a metaphysical one and aims primarily at establishing whether ‘mind’ is something that can be included among natural phenomena or whether, instead, it is an exception. The question about the place of the mind concerns the hypothesis that the mind is just a function of the brain, unless it is actually something more abstract. It is, therefore, a certain functional organization of matter, or if instead of trying to trace the place of the mind it is necessary to look elsewhere, i.e., in social relations, instead of in people’s heads. The third and final question (What is it for?) alludes to the evolutionary advantage that having a mind must entail, given the enormous amount of energy that bodies devote to performing various cognitive tasks. Having a mind represents a significant evolutionary advantage especially when considered in relation to the development of language skills and social cognition.

1.3 Artificial

‘Artificial’ is a word that does not have a good reputation. It is the opposite of ‘natural’, which, by contrast, is a word that meets some of the prevailing sentiments in contemporary culture. ‘Natural’, in fact, refers to anything that is in harmony with the order of the world, from ecology to organic foods. Since contemporary man often feels guilty about the ecological footprint his actions leave on the world, anything artificial seems to allude precisely to man’s inability to realize his own social development in harmony with the rest of the world. Yet ‘artificial’ also has a positive meaning, as it refers to the human ability to produce artifacts through its own intelligence. Technological artifacts, both intangible and cognitive, such as a website, and more concrete, such as a milling machine, are the pride of modem innovation and represent the main evidence of the importance of science in human development.

Technological artifacts are the natural outgrowth of the modem enterprise, driven by scientific knowledge and its vocation to influence social organization in democratic and participatory ways. But if intelligence itself is a candidate for becoming artificial, things change radically. The theoretical conundrum hinges on the fact that modern man sees scientific enterprise and the resulting social change as the result of the development of human intelligence. It is the humanistic project born in the sixteenth century that then animates the birth of experimental science and the Enlightenment that is the basis for typical modern optimism about the future. It is human intelligence that guides the whole process. In what sense could this intelligence become ‘artificial’ without betraying the original humanist project? The answer to this question rests on the confidence that any artificial intelligence should, nevertheless, be no more than an extension of the natural intelligence of man. In the humanistic project of modem science, artificial intelligence is something acceptable only insofar as it is conceived as a prosthesis of human intelligence. For this reason, all theoretical adventures perceived as liberating artificial intelligence from human intelligence are generally viewed with suspicion. Take, for example, Alan Turing’s imitation game, the possibility of an evolution of intelligence that can take us to the Singularity stage, or John Searle’s Chinese room. These are all cases in which we notice a kind of ideological resistance to accepting the very idea of ‘artificiality’ precisely because it seems to deviate from the path pointed out by natural intelligence and thus betray its own mission.

Artificial intelligence is therefore destined to be, at once, the most advanced fruit of the enterprise of modern science, but also the prime suspect for the perversion of its most original spirit. The advancement of artificial intelligence is a socially beneficial endeavor. This is evidenced by the AI4SG (Artificial intelligence for social good) research field, which is an attempt to use artificial intelligence to address social problems and improve the well-being of the world. But, at the same time, artificial intelligence is always seen as something that needs to be monitored and carefully controlled to avoid those apocalyptic scenarios typical of modernity, where machines end up controlling people, thinking is no longer free, and society is reshaped in an authoritarian sense and dominated by obscure technological forces.

1.4 Intelligence

Of all the words we can find in Eega Beeva’s toolbox, ‘intelligence’ is the one characterized by the most general meaning If we refer to the psychology of faculties developed over the centuries of modernity, which in turn is rooted in the way medieval and ancient philosophy had described the totality of human faculties of knowing the world, intelligence appears as a general term that includes all the various cognitive faculties of man Perceiving, speaking, calculating, reasoning, persuading others: these are all ways in which intelligence is articulated in increasingly specific ways. Intelligence is what distinguishes human beings from stupid matter. It is what distinguishes res cogitans from res extensa and would attest to man‘s likeness to God.

The cultivation of intelligence is the main task of man because it distinguishes us from the rest of nature. Hence intelligence requires education and all the social structures like family and school that make it possible. This is why intelligence, if it is in danger of becoming something entirely artificial, is also in danger of betraying its deepest meaning. If intelligence is what makes us human, to make it artificial is to turn it against humans. This way of thinking is still deeply ingrained in many human cultures, especially in the West. But it has also been challenged by numerous scientific achievements. Particularly significant among these are those derived from studies of animal behavior. The Cartesian view of nature had clearly subsumed other animals into mere res extensa. Thus, it was imagined that there was the same qualitative difference between other animals and humans as there is between Dante’s Divina Commedia and the stupidity of a stone we might trip over while walking However, as the study of animal behavior has shown that animal behavior is guided by intelligence and reason just as much as human behavior, the qualitative difference underlying the Cartesian worldview has been strongly questioned.

The possibility that machines are endowed with intelligence is, therefore, one of the most typical adventures of modernity and experimental science. It is, in a sense, its vanguard. If you want to know where the trajectory of modernity is headed, you need to look closely at the trajectory of artificial intelligence. If you like the futuristic outcome of artificial intelligence, then you will like the foundations of modernity. If, on the other hand, the more distant and futuristic scenarios of artificial intelligence evoke a sense of fear and anxiety, then it is precisely the project of modernity that you dislike after all. In that case, you will be drawn to alternative theoretical scenarios, such as postmodernism, the New Age, or some sort of nostalgic return to a historical period in the more distant past.

1.5 Representations

One of the most popular metaphors in cognitive science goes back to Marvin Minsky (1986). It is the idea that the mind is like a society, a vast organized society, something like a city, in which everyone has a job to do and where, on the whole, it seems that everyone cooperates according to a common pattern. If the mind here resembles a city, then its inhabitants are the (mental) representations. The representations are the bearers of meaning because they always have content. And it is precisely because of this content that the mind is a device that does not wander aimlessly, but is something that has meaning. Representations are the bridge between the mind and the world. They are what enable the mind to talk about the world, to be able to relate to it. The mind is, of course, a part of the world. But in order for it to be able to refer to it, it must consist of representations (one could also say, be inhabited by them).

Representations are internal reproductions of particular aspects of the world. They involve the difference between something that is internal and something that is external. If you are a thinker in whose eyes the difference between internal and external, between subject and object, appears as something naive, then representations are not for you. Representations are typical creatures of modernity, a way of thinking that is trapped by the image of a subject trying to make a picture of what is in front of it. Without representations, there is no mind. Or at least there is no mind in the modern sense of the word. Otherwise, the mind can be understood as the ‘breath of the world’, so to speak, i.e., as an emergent property of the world. In this account, the mind is not regarded as a product of the subject, but as a property of the world itself. In other words, the mind is the way the world works. This is a sophisticated idea, one that might meet with the approval of great thinkers like Georg Wilhelm Friedrich Hegel. But it is not the idea on which cognitive science is based. The latter, on the contrary, is based on an idea typical of modernity, emphasized by philosophers like John Locke and Immanuel Kant, according to which the mind is an internal construction of a subject that tries to orient itself in the world by representing it in itself according to a certain logical form.

Representations, however, are not concrete objects. They are like numbers and the basic units of arithmetic. More precisely, mental representations are abstract rules. Their task is to establish a relation between two domains They are comparable to the notion of mathematical function. A function in this context is a relation between two sets that can establish a correspondence between the elements of the first set, called domain, and the elements of the second set, called codomain. In the brain of an animal, a mental representation is like an abstract rule, even if it is physically realized in configurations that are present in the nervous system. The rule provides a link between behavioral types (domain) and environmental influences (codomain) Even if we are able to trace something in the brain that acts as a counterpart to a particular property of the world or object of the world, it is not a mental representation. It is a physical configuration that is connected to a particular class of environmental input via a mental representation. Thus understood, representation has the advantage of being a concept that can be used in a computational context while having a naturalistic counterpart in neural activations and in the class of environmental stimuli that can trigger them. For this reason, representation is the best theoretical construct available to a science of the mind that aspires to be both computational and naturalistic.

This idea of what makes a mental representation is particularly elegant and lightweight. It requires us to represent the phenomenon of knowledge through the naive notion of a subject attempting to create an internal model of the external world through sense systems and the brain. But in other respects it is a particularly liberal idea, that is, it is compatible with most of the theoretical attitudes that have characterized the new science of mind that arose after World War II. It is compatible with an abstract and purely computational idea, typical of the cognitive science of the first decades. It is compatible with a cognitive science that tries to be ecologically correct. And it is compatible with 4E cognition, that is, embodied, embedded, enacted, and extended cognition. Although some radical enactivist philosophers, such as Dan Hutto (1999, 2013) and Shaun Gallagher (2008, 2017), have stated that cognitive science should dispense with the very idea of representation, this is actually an overstated conclusion. They object to the idea that the mind consists of manipulations of explicit representations characterized by the propositional format. They also try to counter the view that knowledge has to do with the mirror metaphor, i.e., that it is essentially about creating a reproduction of the world. On the contrary, they think (rightly) that knowledge is primarily to do with the notion of ‘action’. The mind is not a speculation of the world to which action is then e...