Recently, a book appeared in France and immediately received a full-page review in a major German newspaper: La fin de l’individu. Voyage d’un philosophe au pays de l’intelligence artificielle by the French philosopher Gaspard Koenig. The title says it all: a philosopher’s voyage into the land of artificial intelligence merits the conclusion that the digital society is the end of the individual person as we know it. However, does one have to accept this conclusion? Is digitalization indeed the end of the autonomous person and of free will? This question, recently discussed not only by a prominent French philosopher but also by many other voices from philosophy, sociology, cultural theory and computer science, is obviously important for many legal issues as well. It raises the question of how legal debates can make use of the vast treasure of knowledge generated by the whole array of sciences and humanities. This is where legal theory comes into play.
Law as a Social System
The first step towards an answer on how to fuse legal and social knowledge into legal theory is to understand the law as a subsystem of modern society. Any major social change such as digitalization is highly likely also to affect the law, in particular, its conceptual structure. How can legal theory conceptualize such a shift? One useful way is to look at key concepts of modern law such as autonomy, and related concepts such as the individual, the person and freedom of will. All of these share the property of being not only legal but also social concepts with a strong philosophical foundation that bears the normative basis of modern society. Internal shifts of meaning of these and other core concepts of modern law thus provide a litmus test for similar changes within the normative basis of modern society.
In order to understand such changes and to formulate an adequate legal theory of the digital society, however, it is important to look not only at the outward surface of a concept like autonomy but also at the more subtle change of its internal meaning against a rapidly changing social background. ‘Autonomy’ has been a legally relevant concept for at least the past two centuries. Thus, in order to ask a theoretically meaningful question about the impact of the digital society on legal autonomy, it does not lead very far to ask whether the law still recognizes individual autonomy in the digital age and to answer that question with a simple ‘yes’ or ‘no’, as Koenig seems to suggest. In fact, the answer to that question seems to be a clear ‘yes’ without teaching us anything about the specific form that autonomy takes on under the changing social parameters of the digital society. Instead, it seems more useful to ask questions like what autonomy specifically means under the changed social circumstances of the digital society, whether these changes cause new disruptions in the philosophically loaded normative meaning of autonomy as a legal concept, and what all of this might teach us about our self-definition as autonomous agents within the digitized modern society.
The Modern Person: Individual or Dividual?
What does it mean to be an autonomous person? While this question is a legal issue, it is also a fundamental problem of philosophy. More precisely, the legal import of the question depends on the philosophical background of the concepts of autonomy and the person. Thus, a sufficient understanding of the latter concepts is essential for coping with the former question of law. In Western thought, the specifically modern understanding of the person emerged in the 17th and 18th centuries with the rise of enlightenment thinking. One of the earliest thinkers to explore the idea of the moral person as the bearer of natural rights was Samuel Pufendorf. Writing one century before Immanuel Kant, Pufendorf already explored the idea that there is a quality of freedom to act morally attached to every moral person and that it is this very quality which bestows equal human dignity upon each and every human being. This idea reached its final and most elaborate form in the writings of Immanuel Kant around 1800.
This is, however, where things became difficult. In the history of ideas, the year 1800 also demarcates what is known as the Sattelzeit at the transition to industrial modernity. The Sattelzeit describes the formation of the modern, differentiated society which has emerged since the 19th century until today. One of the core differentiations of modern society, unknown to premodern conceptions of society, is the divide between law and morality. This distinction implies, in particular, a separation between the moral concept of autonomy and the legal concept of freedom. Even though the terms are often conflated and used interchangeably in morality and law, they do not mean the same thing. Interestingly enough, this separation occurs at the same time as autonomy appears as a core concept of modern law.
In Kantian philosophy of law, autonomy means internal freedom, which is equivalent to freedom of conscience as a moral category, ruled by the categorical imperative. The legal concepts of freedom or ‘autonomy’, on the other hand, pertain to the sphere of external freedom, also known as Willkür. There are, therefore, two complementary concepts of freedom that point in opposite directions: one inward towards one’s own conscience, and the other outward towards the other persons. Internal freedom is equivalent to moral autonomy, whereas the relationship between autonomy and external freedom is more complex. According to the Kantian principle of right, ‘any action is right if it can coexist with everyone’s freedom in accordance with a universal law, or if on its maxim the freedom of choice of each can coexist with everyone’s freedom in accordance with a universal law’. Thus, Kant tries to address the problem of conflict between concurring external freedoms with a formula closely resembling, or actually paralleling, the famous categorical imperative. The latter, however, only applies to the internal spheres of morality. Kant’s concept of right thus leads to a normative dilemma for the external conception of law embraced by modern society: the law, restricted to the external sphere of freedom, cannot command moral action, because the law holds no power over the internal forum of morality. Nonetheless, the law cannot do without the tacit presupposition of moral action, because the Kantian principle of right is only functional if human beings use their external freedom responsibly, that is, in analogy to the moral principle of right as guided by the categorical imperative. The Kantian gap between law and morality, between external and internal freedom, thus presumes a blank cheque of social morality in order to work as a social theory.
Against this background, we can put the key question with regard to autonomy in the digital society as follows: in what sense does the digital society change our self-understanding as autonomous agents, and how does the law reflect this change when it addresses the autonomous person as a legal actor in the sphere of external freedom? In other words, is there a restricted sense of moral autonomy in the digital society, and does this have any effect on the construction of the person as the bearer of legal autonomy? That there is in fact some change taking place in our self-definition as autonomous agents under the conditions of digitization seems indubitable. This is why a title like Koenig’s ‘end of the individual’ hits such a nerve in the current discourse. What exactly, however, is this change all about? Is it a truly qualitative change, a digital disruption, as thinkers like Koenig seem to suggest, so that we do indeed have to cope with something entirely new beyond the old paradox of Western individualism as exemplified by Kant’s inability to bridge the gap between legality and morality? Or is the rise of the digital age just the next step in the development of structures already inherent in modern society, namely, the advent of yet another network technology with—of course—dramatically increased potential compared to previous mass communication media such as the telephone or letterpress printing, but without the extreme disruptive potential ascribed to it? If the latter thesis holds true, the underlying structure of the digital society might not be so different from the pre-digital society after all.
I argue that the second theory in fact offers more fruitful insights than the only superficially ‘radical’ first theory that the digital age comes as something essentially disruptive to our society—like a biblical plague. Digitalization is not a biblical plague. Rather, modern society has been proto-digital from its very beginning. Its structure is essentially digital. Digital methods have, therefore, been a part of social practices at least since the 19th century, when the potential of large-scale social planning and social statistics became evident. Thus, the tools of the digital world amount to no less—but also no more—than the structurally congenial tools of modern society, the fulfilment of the technical promise of modernity, which is essentially countable, in a word: dividual.
The concept of the countable ‘dividual’ as the core of modern personhood was aptly coined by Gilles Deleuze as early as 1992 as the antonym to the old individual associated with indivisible autonomy in the classic sense. According to Deleuze, ‘we no longer find ourselves dealing with the mass / individual pair. Individuals have become “dividuals”, and masses samples, data, markets, or “banks”.’ Digitization, in other words, has the potential to transform the indivisible individual into a divisible dividual. The individual person as an impartible entity endowed with human dignity has become obsolete as a conceptual tool of self-description within the technical framework of the digital society. Digital technology makes it possible to circumvent exactly the indivisible character of the person as a social construction in all kinds of contexts, because it is now possible to generate a better, more precise representation of the empirical human being as the product of digital information. Digitization apparently poses a fundamental problem for our self-understanding as autonomous persons. We can only be moral persons if we are individuals gifted with indivisible autonomy. If the indivisible quality of the person disappears behind algorithmically generated type profiles, this means that the individual is no longer conceived as indivisible, equal and free, but rather as divisible, calculable, predictable and in that respect unfree.
As a consequence, the concept of individual right, the unquestionable basis of modern law, seems to collapse into a ‘dividual’ right as well. The concept of right seems to be in danger of no longer addressing human beings as indivisible, free and autonomous persons, but rather as the predictable results of mere statistical distributions and type profiles. The individual disappears behind big data and algorithmic readability. The concept of personhood becomes irrelevant for the purpose of legal control, because data statistics literally know the individual person better than she knows herself. Moreover, big data analysis offers the astonishing feature that it is no longer dependent on the personal data of any given individual. It is in fact possible to predict the statistically relevant behaviour of any individual with greater precision from mere data patterns than from the individual’s behaviour itself. All it takes is to combine the statistical frequency distribution of the behaviour of classes of given individuals with the skilful application of search criteria.
Is this Gaspard Koenig’s ‘end of the individual’? No, it is not. If it were true that we have been indivisible individuals up to now and that we are losing this individuality, we would most certainly fight to retain our self-definition. But that is obviously not what is happening. Instead, we are witnessing the opposite. Perhaps the most astonishing aspect of the development of the digital society is that we actually want it to happen exactly the way it is unfolding. We reinforce it every day by voluntarily consenting to giving away our data with every move we make on the internet. Everyone with a smartphone or those active on social media or in the digital economy actively participate in building a digital society on the basis of personal data without any certitude about how or in what contexts this data can and will be used in the future. At this point, virtually all of us already have a double existence in this world, as human beings and as data avatars. The fact that we actually want this to happen is the true challenge for our thinking about autonomy. For we assume quite naturally that we continue to exist as indivisible persons who are capable of exercising norma- tively significant autonomy in the sense of Kantian philosophy, whereas in fact we have stopped existing as individuals in the data world long ago and just have not noticed it yet. One can frame this as a paradox: what does it teach us about our understanding of autonomy that we believe ourselves to be fully autonomous also and especially with regard to the heteronomous predictability of our own use of autonomy?
The Digitally Conditioned Self at the Crossroads of Legal Theory
I will conclude by arguing that I do not believe in the end of the individual, but I do believe that we have to get used to a novel concept of autonomy in the digital world, whether or not we approve of it. Again, the starting point is that digital networks are not so different from pre-digital networks or communication media after all. Human beings have always depended on networks to define themselves. They cannot express themselves in any other way than through interaction with others. This applies particularly to the technical networks of the industrial age. It is, therefore, unreasonable to imagine a person endowed with full autonomy existing prior to and outside any network and then turning towards the network in order to say ‘yes’ or ‘no’ to their participation. The reality is quite different: we can only unfold our autonomous personality through participating in networks. We have to define autonomy also as a product and not only as a limit of our development through networks. There is no such thing as uncorrupted autonomy, defined as network-free autonomy.
What in particular, then, makes the digitally conditioned self so dubious to critics like Gaspard Koenig? As argued before, it is the paradox of modern society that it offers external freedom but cannot, at the same time, demand the exercise of moral autonomy in its fulfilment. And it is precisely this phenomenon, this built-in deficit of modern society, that is amplified and exploited by digital mass media and especially by social networks in an unprecedented way. Social networks seem to have developed into a congenial infrastructure for antisocial behaviour. By using social media, the digital society is deliberately conditioning itself into a dysfunctional use of freedom which runs against its tacit precondition of moral autonomy. If this is the case, however, why does the law not simply prevent abuses of digital freedom by appropriate regulation? The answer to this question, again, turns on the philosophical dilemma of autonomy underlying the legal concept of freedom: it is not the case that we do not regulate digital platforms because we cannot do so, but rather because we do not want to do so. In more precise terms, we cannot want to do so, because the notion of undivided external freedom by definition prevents us from selecting the desired from the undesired uses of legal external freedom.
At this point, one may still argue that each individual nonetheless remains free to exercise moral autonomy by using digital technologies responsibly. There is, however, a second, even more serious side of the problem concerning the form of freedom into which we condition ourselves in digital contexts. I describe this form as a collective education towards a consumptive use of freedom. Freedom itself becomes the object of consumption in the way it is made accessible through digital tools. In other words, the digital infrastructure we have built teaches us to use external freedom like a commodity and thereby to unlearn the use of moral autonomy. The reason for this is a formal—one might even say architectural—problem: the consumed freedom remains passive by virtue of its very form, which is conditioned through the one-sided, passivising architecture of the digital infrastructure which virtually reduces the world to a user surface.
This idea is, finally, also the key to overcoming the general complaint of the loss of personhood in the digital age as voiced by Koenig. We should shift the power of critique from the concept of personhood to the architecture of the digital infrastructure. The important thing to note is, again, that there is no freedom and no autonomy in the technical world untouched by its preconditions. The crucial question of digital autonomy is, therefore, not whether to use digital technology at all, but rather how to articulate autonomy within the structures of the technically preconditioned world—and, even more importantly, how to build a technical infrastructure which leaves the individual pathways to autonomy intact. Algorithms have the power to deconstruct individual human beings into statistical artefacts. But neither power nor technology are evil as such, and human beings have been calculable all along. The true problem for the digital self lies in the obstruction of per- sonal autonomy by conditioning individuals into the merely passive exercise of freedom. This is where the regulation of the digital world should begin, that is, with the philosophical demand for autonomy remaining in the background.