Preface

 It is, and indeed has been, widely acknowledged through the ages that certain roots of modern scientific inquiry can be traced back to the work of Aristotle. This fact strongly supports the view that Aristotle’s thought retains an enduring relevance, at least in part. On the one hand, several aspects of his thinking –by no means negligible in scope or depth– remain valid and continue to occupy a prominent place of classicism in the “Pantheon of Sciences.” On the other hand, in those domains where Aristotle’s work has been surpassed –an entirely natural development over the course of centuries, especially under the rigorous scrutiny of scientific theory through experiment and observation, specifically in the field of the exact sciences–, it still holds its place within the field of the History of Science, as acknowledged by many distinguished scholars in the relevant disciplines. This emblematic relevance of Aristotle’s thought is further affirmed by the fact that the Nicomachean Ethics is widely taught in the scientific field of the Humanities internationally, especially in many Philosophy departments of the American universities.

A. Based on these observations, it is far from paradoxical to argue that Legal Science, as part of the Theoretical Sciences, owes much to Aristotle’s theoretical inquiry. In particular, the Nicomachean Ethics, the Rhetoric, and the Athenian Constitution contain analyses that remain extremely useful to this day. These works, through a theoretical approach which employs a rigorous scientific methodology, especially with regard to Law and Justice, demonstrate clearly the essential and multifaceted debt contemporary Legal Science owes to Aristotle’s oeuvre. More specifically, the study of Aristotle’s work on Law and Justice continues to contribute, significantly, to our understanding, on the one hand, of the Rule of Law as a cornerstone for the legal structure in general, and, on the other, of Equity as a fundamental principle within the overall system of Justice for the interpretation and implementation of the law in its full normative scope. It is worth noting that international scholarship, as will be examined in some of its aspects below, provides unambiguous testimony to this.

B. Furthermore, if we consider the most recent developments in Artificial Intelligence, it is easy to conclude that Aristotle’s views on the quintessence of Law and Justice –and, thereby respectively, on the Rule of Law and Equity– are also highly pertinent to defining the boundaries of the use of this emblematic form of modern technology within the aforementioned field of Legal Science. This is all the more significant, given the global tendency –particularly evident in the United States– which is quite alarming as to the Rule of Law and the Principle of Legitimacy within the institutional framework of Representative Democracy, toward the reckless use of Artificial Intelligence, both in the formulation of legal norms and in their practical application within the context of the administration of Justice in concreto (with reference to certain facts or cases). The foregoing considerations outline the content of the two sections that constitute the brief study presented below (cf. my articles, “Observations on the Contemporary Relevance of Aristotle’s Theory of Equity in the Field of Modern Public Law”, Administrative Law Review, vol. 4, 2021; and “‘Dilemmas’ of Legal Science in the Framework of Artificial Intelligence”, Administrative Law Review, vol. 6, 2023). The first part offers, in summary, Aristotle’s views on Law and Justice, with an emphasis on the Rule of Law in the former case, and the principle of Equity in the latter. The second part provides a concise demonstration that Aristotle’s aforementioned views can serve as a valuable guide in the effort to restrain the reckless use of Artificial Intelligence –both in the shaping of legal norms and in the exercise of judicial authority by the members of the Judiciary– guided by their institutional and normative purpose of defending Law and Justice. Ultimately, this defense extends to the very institutions of Representative Democracy, and therefore also to Freedom and to the Fundamental Human Rights.

I.  Aristotle’s Philosophical–Legal Thought on Law and Justice and Its Implications for the Normative Structure of the Rule of Law and of Equity

Anyone examining the legal coordinates of Aristotle’s philosophical–legal thought, based primarily on the Nicomachean Ethics and the Rhetoric, is easily led to the evidently valid conclusion that, for the Stagirite, Law, as a normative system of deontological content, is enacted, interpreted, and implemented with the principal aim of ensuring the cohesion of the social body governed by its provisions. And this aim can only be achieved under the conditions of Justice. Thus, employing the terminology of contemporary legal theory, one may assert that, according to Aristotle, legal rules, by their very nature, ought to be interpreted primarily teleologically – that is, in accordance with the purpose that led to their enactment. This purpose is inextricably linked to the establishment of Justice, understood in its deontological form, as a self-evidently indeterminate legal concept that is specified in each case –unlike indeterminate evaluative concepts– through the interpretation of the applicable legal rules. This, in turn, implies that there is no real discretionary power on the part of the interpreter and applier of the Rule of Law in this process of specification, since such discretion is recognized, from the standpoint of legal science, only in cases involving indeterminate evaluative concepts, and only insofar as they are subject to normative regulation through legal rules.

A.  Justice according to Aristotle

In Aristotle’s thought, Justice is a philosophical Janus. This is because it presents itself in two aspects which, by their very nature and despite their conceptual differences, are interrelated and mutually reinforcing:

1.Justice as a Virtue

First and foremost, Aristotle conceived of Justice in its generality, and more precisely, as a Virtue. Yet it is a Virtue not to be understood solely in theoretical terms or as a matter of normative imposition. Quite the opposite: for this reason, it cannot, in any sense, be subsumed under the category of indeterminate legal concepts.

a) Justice as Virtue, in the Aristotelian sense, primarily functions as a guiding principle for human behavior in everyday life. The citizen is expected –on grounds related to the achievement of harmonious social coexistence and, subsequently, of social cohesion– to act virtuously. The dimensions of such virtuous behavior are defined by the specific imperatives –or we might say the specific directives– of Justice.

b) At its philosophical zenith, according to Aristotle, this first aspect of Justice acquires the characteristics of a perfect Virtue. This occurs when Justice reaches that completed stage which both synthesizes and presupposes the other particular virtues – moral, social, and political. Notably, political virtues, as elements of human behavior, are inextricably linked –indeed, almost identified– with the observance of Legitimacy and respect for Equality, understood in its proportional sense, as will be clarified below.

2. Justice and Proportionality

Secondly, Aristotle examined and defined Justice, not only in its purely philosophical essence mentioned above, but also in institutional and political terms, understood lato sensu. This perspective aligns perfectly with the likewise lato sensu normative nature of Justice as an indeterminate legal concept—more precisely, as a principle that imparts to individual legal norms the foundational characteristics of Proportionality, which originates ultimately in the foundations of Harmony.

a) This inquiry led Aristotle to his conception of the proportional aspect of Justice, which is identified –at least in many aspects– with its “twin sister,” proportional Equality. This form of Equality presupposes, as a general condition for the formation, interpretation, and application of legal rules, an equal treatment of essentially similar situations and an unequal treatment of essentially dissimilar ones. This is because the notion of Proportional Justice, with proportional Equality at its forefront, is by definition opposed to the unequal treatment of equals and, conversely, to the equal treatment of unequals. Both practices inevitably result in a form of contradictio in adjecto with respect to the substance of proportional Equality and, consequently, proportional Justice. In this domain, as will be discussed in greater detail below, Aristotle’s general principle of Equity also plays a key role, by facilitating the proper interpretation and application of legal norms so as to ensure, among other things, the aforementioned equal treatment of essentially similar cases and the unequal treatment of essentially dissimilar ones.

b) Furthermore, and as a consequence of the foregoing, Aristotle’s concept of Proportional Justice—through the application of proportional Equality—can be distinguished, based on its specific modes of implementation, into the following forms:

b1) Distributive Justice. This is the form of Justice whose observance ensures, in practice, both the fair distribution of goods among the members of a given society and the discernment of the differences between them—where and when necessary—in terms of proportional Equality.

b2) Corrective Justice. This form of Justice makes the restoration of proportional Equality possible, whenever and to the extent that it has been disrupted in practice. As is readily understood, in such cases, Aristotle’s Corrective Justice operates through the philosophical and institutional means of interpreting and applying legal rules which no longer follow principles of geometric but rather of arithmetic proportionality. It is this form of proportionality that underpins today, for example, the organization and operation of compensation mechanisms, whether in the field of Public Law or Private Law.

b3) And Retributive Justice. Despite the terminological indication, this form of Justice bears no relation to the principle of lex talionis (the law of retaliation), as it surfaces in Criminal Law theory. Quite the contrary: Retributive Justice refers to the proportional reciprocation of goods and services, aiming at the full realization of proportional Equality within the framework of legally governed daily transactions, as dictated by social necessity and, ultimately, the safeguarding of social cohesion in terms of fundamental solidarity.

B. A Compelling Demonstration of the Contemporary Relevance of Aristotle’s Thought on Justice

One need not labor to grasp the extent to which Aristotle’s aforementioned views on Justice—particularly in their proportional dimension, as outlined in the philosophical and institutional essence of proportional Equality—remain timely. It suffices to point to the magnum opus of perhaps the most prominent contemporary representative of philosophical reflection on Justice in the United States, and more broadly, in the West. Specifically, to John Rawls’s A Theory of Justice (The Belknap Press of Harvard University Press, 1971; Greek edition: Polis, 2010). In broad terms, Rawls “encounters” in his aforementioned work—notwithstanding the reservations expressed by some of his commentators as to the depth of this encounter—Aristotle’s thought, especially in the context of Distributive Justice, as already explained. More precisely, John Rawls conceives of Justice as a foundational principle within the Legal Order, through which individual legal rules are interpreted and applied under conditions approximating ideal fairness. This ideal is practically ensured through the respect for and internalization of the imperatives of proportional Equality.

1. John Rawls on Justice

In this regard, John Rawls understands and analyzes Distributive Justice always in light of the principle of proportional Equality:

a) First, as equality in the exercise and enjoyment of Freedom in all its institutional and political expressions. Here, Rawls essentially begins from the very foundations of Representative Democracy, which—in its modern and most normatively recognized form—is conceived as a framework for the organization and operation of institutions that guarantee the exercise and enjoyment of Freedom, as this acquires substantive meaning and content through the unimpeded exercise of Fundamental Human Rights. These are guaranteed not only by the Constitution—as the supreme norm of every Legal Order—but also by International Law in its various branches, and by European Law, within both the European Union and the legal systems of its Member States.

b) Second, as equality which, in light of the principles of Distributive Justice, establishes and ensures in practice equal opportunities for each rights-holder to exercise their rights effectively. This form of equality must result in the creation of a status of initial equality – a common starting point for every rights-holder in order to exercise their rights effectively, in accordance with the fundamental imperatives of Distributive Justice. Such a condition is absolutely necessary, for without the guarantee of equal opportunities and, consequently, without the corresponding creation of the basic conditions of equality, the exercise of rights is practically nullified. It becomes vulnerable to the corrosive toxicity –both social and economic– of the ever-widening inequalities of our time. This is felt not only in the sphere of the standard individual defensive rights but foremost in the field of social rights and the Social Rule of Law in general. The point becomes even clearer when one considers the reciprocal relationship between rising inequality and the decline of the Social Rule of Law: the more inequalities increase, the more the structural elements of the Social Rule of Law are undermined; and conversely, the more the Social Rule of Law is weakened –especially at the current, quasi-geometrically escalating pace– the more the scale of inequality extends to the point of social destitution.

 2.The “Encounter” between John Rawls and Aristotle

In light of these considerations, amid the current dystopian reality of the declining trajectory of Representative Democracy, driven chiefly by the aforementioned rapid widening of inequalities,  the ensuing fundamental undermining of the Social Rule of Law, and the dangerous societal outlook heralded by uncontrolled Technological Unemployment (a symptom which starkly illustrates the risk of unemployment becoming endemic in our societies when, as John Maynard Keynes reminds us, the cornerstone of a healthy capitalist system is the assurance of full employment as much as possible), John Rawls can be said to confirm certain core Aristotelian positions on Distributive Justice, grounded in the idea of proportional Equality. This is evidenced by the following:

a) In the exercise of Fundamental Human Rights and for the enhancement of social cohesion, there exists a fundamental rule –already stated above– regarding the structure of Equality in its strictly proportional sense: equal treatment of essentially similar cases and unequal treatment of essentially dissimilar ones. The application of this rule presupposes –among other things, but above all– the enactment of specific legal norms and the adoption of appropriate practices aimed at the gradual reduction of inequalities and the effective reconciliation of conflicting positions within each social body. The goal is to avoid the rupture of the social fabric and to secure, as far as possible, the triumph of the idea of Justice. In this context, the role of the principle of Equity—as articulated by Aristotle and further elaborated below—is of decisive importance in interpreting and applying legal norms aligned with the ideal of Justice.

b) It is therefore neither surprising—nor, by consequence, accidental—that the scholarly community today widely accepts the following: In evaluating the work and contribution of John Rawls with respect to the philosophical, institutional, and political conception of Justice—and especially through the analysis by which he developed his so-called “Theory of Justice as Fairness,” particularly through the harmonization of divergent views to reach their necessary unity so as to avoid endangering social cohesion—we are led to the conclusion that: At least in this dimension of his work on Justice, Rawls’s theoretical reflection is closely aligned with Aristotle’s dialectical and, more broadly, methodological approach. Let us not forget that this dialectical method of Aristotelian thought survived to the present thanks to the substantial influence of Thomas Aquinas, who—according to historically irrefutable evidence—became familiar with Aristotle’s work through the invaluable prior engagement with it by John of Damascus, a leading figure of Byzantine letters and, by extension, of Byzantine civilization.

C. Equity as a Fundamental Condition for the Application of the Principles of Distributive Justice in Aristotle

From what has already been pointed out –even if only by implication– it is evident that Aristotle’s overall conception of Equity is inextricably linked with his broader theory of Justice. In particular, it is clear that one of the most crucial means –and arguably the most essential institutionally– for the realization of Justice, especially in its distributive (and, therefore, corrective) form, is Equity. Equity functions as a guiding principle both for the formation of legal norms, which govern the daily lives of the members of a society, and for their eventual interpretation and implementation. Consequently, as a decisive “companion” and “facilitator” in the fulfillment of Justice, Equity constitutes, by synecdoche and in a quasi-automatic fashion, an indeterminate legal concept, as described earlier. This indeterminate nature is of paramount significance, particularly for judicial bodies in the broad sense of the term, when exercising their adjudicative functions.

1.The Pre-Aristotelian Conception of Equity

Preliminarily, it should be clarified that the concept of epieikeia (Equity), within the framework of philosophical reflection in Classical Greek Antiquity, was not first examined by Aristotle.

a) This is because certain philosophers prior to Aristotle had already focused—directly or indirectly—on the idea of epieikeia.

a1) A characteristic example is that of Democritus, as preserved—since we know very little directly from the vast corpus of the philosopher from Abdera—by Stobaeus (Diels-Kranz, Die Fragmente der Vorsokratiker, 1992, II, 195, 252 [134N]). According to Stobaeus, Democritus expressed the following thoughts on epieikeia: «Τά κατά τήν πόλιν χρεῶν τῶν λοιπῶν μέγιστα ἡγεῖσθαι, ὅκως ἄξεται εὖ, μήτε φιλονικέοντα παρά τό ἐπιεικές μήτε ἰσχύν ἑαυτῷ περιτιθέμενον παρά τό χρηστόν τό τοῦ ξυνοῦ». This means that, for Democritus, above all other obligations related to the organization and functioning of the City-State, one must prioritize those that, on the one hand, ensure effective governance free from strife and discord which undermine epieikeia, and, on the other hand, those that prevent the obtainment of excessive power, which is by nature opposed to the common interest and, consequently, the common good.

a2) Another example is offered by Gorgias of Leontini (Diels-Kranz, op. cit., II, 285, 6 [15–16]): «Πολλά μέν δή τό πρᾶον ἐπιεικές τοῦ αὐθάδους δικαίου προκρίνοντες». Here, Gorgias—clearly influenced by the Eleatic School, especially Zeno and Parmenides—emphasizes on the role of epieikeia in restoring the essence of dikaiosynē (Justice). He concludes that, to this end, the gentle moderation of epieikeia, which serves the common good, should be preferred over the rigid literal application of legal norms. This maxim of course may verge in practice on the brink of intolerable injustice. This conclusion was later concisely articulated by the Roman legal thought in the maxim summum jus, summa injuria, which simultaneously highlighted the futility of rigid absolutism in achieving Justice in practice, as expressed in the aphorisms, dura lex sed lex and the much later fiat justitia et pereat mundus.

b) What must therefore be credited to Aristotle’s thought in this regard—as will become clearer in the subsequent analysis—is that he developed the concept of epieikeia philosophically in a far more comprehensive manner, adding to many of its institutional and political dimensions.

b1) And he did so with such methodological rigor that he elevated it to a quasi-perfect level, especially considering the philosophical context of his time. This accomplishment becomes even more impressive when we consider that other philosophers, including some of the most prominent of his era, expressed serious reservations about the idea that epieikeia could and should be used even as a corrective instrument for the interpretation and implementation of established legal norms.

b2) Within this context, Aristotle, in formulating his theory of epieikeia, did not hesitate to enter into a kind of intellectual conflict with the thought of his teacher, Plato. In the Laws (757d–757e), Plato had observed: «Τό γάρ ἐπιεικές καί σύγγνωμον τοῦ τελέου καί ἀκριβοῦς παρά δίκην τήν ὀρθήν ἐστιν παρατεθραυμένον, ὅταν γίγνηται». This means that, for Plato, the invocation of epieikeia—and the tolerance it entails—against the literal and strict meaning of legal norms in their practical application, risks producing a kind of fracture in the proper administration of Justice.

2. The Concept of Epieikeia in Aristotle

Broadly speaking, Aristotle’s thought on epieikeia (Equity)—as a principle that functions to support the restoration of the idea of Justice, primarily in its distributive and, consequently, in its corrective conceptual form—covers a very wide spectrum. This spectrum extends, roughly, from the reasons which render epieikeia both appropriate and necessary for the proper administration of Justice, to the methods through which it can and ought to be employed in practice for the interpretation and implementation of legal norms, always with the purpose of dispensing Justice according to its inherently philosophical and institutional purpose. It is important to note at the outset that, among these methods, Arbitration (diaitēsia) holds a noteworthy place in Aristotle’s thought—i.e. the adjudicatory process which occurs outside the formal framework of Justice as administered by the officially authorized judicial bodies of the state.

a) It must first be clarified within which broader philosophical category Aristotle places epieikeia, as this decisively influences his further determination of the notion. The issue for Aristotle was that, e.g., in Plato’s account—as noted above—epieikeia could not be classified, without qualification, among the specific concepts encompassed by the broader notion of aretē (Virtue). This is because, by its nature, it contains traces of weakness or retreat from the full realization of Justice in its practical application:

a1) In response to these theoretical hesitations of Plato, Aristotle is philosophically clear—particularly as regards the evidence he provides in the Nicomachean Ethics. When he states, for example, «ἡ δὲ ἀρετὴ (…) τοῦ μέσου ἂν εἴη στοχαστική» (Nicomachean Ethics, 1106b 15–17), thus returning to the roots of Presocratic reasoning, he quite evidently defines Virtue in terms of a balance between two extremes and the consequent imperative of harmonizing opposites wherever necessary. In other words, he defines Virtue as a mean (μέσον) for responding to extreme conditions, which are disruptive to social cohesion and harmony. Within this framework, Aristotle’s thought unequivocally classifies epieikeia among the particular forms of the general category of Virtue — that is, the form of Virtue pertaining to human behavior when such behavior must conform to enacted rules of law, established by all legitimate means.

a2) And this is because epieikeia, by its constituent features—more clearly elaborated below—serves, almost by definition, the principle of the mean (mesotēs), particularly when it ultimately contributes to the practical realization of Justice as a whole. It does so through channels that reflect the more specific expressions of Justice, especially the Distributive and, consequently, the Corrective Justice. More specifically, epieikeia in Aristotle, as the “steward” (therapainis) of both Distributive and Corrective Justice, facilitates—through its inherent moderation—the selection of those interpretative or applicative solutions to legal norms which prevent the de facto engendering of extreme situations that are entirely contrary to the essence of Justice in globo.

b) The aforementioned conclusions concerning the philosophical conception of epieikeia as a Virtue are further confirmed and specified by Aristotle’s subsequent positions—this time regarding its primarily institutional function within the field of interpretation and application of legal norms:

b1) According to Aristotle—as indeed more broadly in ancient Greek philosophical-legal thought—the law must have, by its very nature, a general and indeterminate content. Law, therefore, cannot be individual in nature, that is, it cannot regulate one or more specific legal relationships. This is because an individual law, intrinsically and inevitably, leads to a violation of the principle of proportional Equality, potentially resulting either in the unequal treatment of essentially similar cases or in the equal treatment of essentially dissimilar ones. Consequently, individual law may end up violating the very essence of Justice, especially in its distributive form. On the other hand, the necessarily general and indeterminate nature of law –and therefore the legislator who enacts it– cannot possibly determine with absolute precision the full scope of regulatory coverage that law is intended to provide, since it is beyond human capacity to foresee, in advance, every particular case that may arise in the future within the relevant domain. The task of specifying the general and indeterminate nature of law and of addressing its normative gaps, thus, belongs, in Aristotle’s view, to the adjudicatory function, i.e. to the judge: «Ὅταν [οἱ νομοθέτες] μή δύνωνται διορίσαι, ἀλλ‘ ἀναγκαῖον μέν ᾖ καθόλου εἰπεῖν, μή ᾖ δέ, ἀλλ‘ ὡς ἐπὶ τὸ πολύ» (Rhetoric, 1374a 30–31).

b2) In order to fulfill this role successfully, the judge must also invoke epieikeia as a principle for the interpretation and implementation of the legal rule—a principle which excludes extreme solutions for addressing various kinds of regulatory gaps, which are, by definition, incompatible with the concept of Justice. Aristotle illustrates this point with the following example: A general and indeterminate law cannot, for instance, define, in advance or sufficiently, when a certain person truly commits an injustice by striking, in any way, another person with his hand. There are cases in which the vagueness and generality of the law, when interpreted strictly and narrowly, may lead to the conviction of someone whose behavior is not, in substance, unjust. Here, epieikeia intervenes to assist interpretation or fill in the normative gap of the law, so as to facilitate the correct administration of Justice: «Κατὰ μὲν τὸν γεγραμμένον νόμον ἔνοχός ἐστι καὶ ἀδικεῖ, κατὰ δὲ τὸ ἀληθές οὐκ ἀδικεῖ, καὶ τὸ ἐπιεικὲς τοῦτό ἐστιν» (Rhetoric, 1374a 36–1374b1).

b3) Subsequently, Aristotle, in accordance with the above and based on his view that epieikeia belongs to the broader field of Virtue, proceeds to more specialized reflections on the concept and function of epieikeia within the framework of adjudicating Justice. In particular, he offers analyses that define –with remarkable, at his time, clarity and insight– the relation between written law and epieikeia as a principle for the interpretation and implementation of the former. A few isolated remarks from Aristotle suffice to “confirm the reliability of the argument”: Aristotle begins from the philosophical–legal premise that dikaion (the just) and epieikes (the equitable)—and hence both written Law, such as statutes, and Equity—must function harmoniously in the interpretation and implementation of legal norms, with the goal of achieving a just outcome. Accordingly, both the just and the equitable are equally important—indeed crucial—for the interpretation and implementation of law. However, in the end –according to Aristotle– what is decisively important, and in many cases crucial for the proper administration of Justice, is epieikeia.

b4) In other words, here the Stagirite demonstrates a kind of preference in favor of the principle of epieikeia in the interpretation and implementation of legal norms, as opposed to the absolute and rigid literal expression of the law that enacts them. He justifies this preference by explicitly accepting, expressis verbis, «ὅτι τὸ ἐπιεικὲς δίκαιον μέν ἐστιν, οὐ τὸ κατὰ νόμον δέ, ἀλλ᾽ ἐπανόρθωμα νομίμου δικαίου» (Nicomachean Ethics, 1137b 11–13). That is, he affirms that the equitable corresponds to the idea of dikaiosynē (Justice), not in accordance with the letter of the enacted law, but rather as a corrective mechanism of the certain law, whenever and to the extent that this is required in order to restore its provisions—so that, in their interpretation and implementation, they may fully reflect the true concept of Justice in its mission to effectively adjudicate, as intended. Thus, for Aristotle, the principle of epieikeia is indeed included within the general framework of Law. However, it is not absolutely identical with the enacted written law but may—and ought—to be invoked when that law, due to its gaps or regulatory shortcomings, must be corrected in terms of its absolute generality through appropriate interpretation. In this way, the law can more fully correspond to the entire set of imperatives expressed by Law and Justice.

b5) On the basis of these observations, Aristotle reaches—albeit fully aware, as his entire work shows, that omnis definitio periculosa est— a kind of philosophical–legal definition of epieikeia which remains classic even today: «Ἔστι δὲ ἐπιεικὲς τὸ παρὰ τὸν γεγραμμένον νόμον δίκαιον» (Rhetoric, 1374a 27–28). Some authors have argued that, in this definition, Aristotle ultimately sets epieikeia in opposition to written, enacted Law. However, this interpretation is mistaken and reveals a fundamental misunderstanding of Aristotle’s authentic view in this matter. In reality, Aristotle—consistent with all his prior analyses on this subject—summarizes them with impeccable logical precision, clarifying that the equitable is not the opposite of the norms of the written law. Rather, it may follow a parallel path to them in defending Law and Justice, and may even go beyond them, insofar as this is necessary for realizing the true meaning of both Law and Justice.

b6) Using the tools of modern legal terminology, it is both feasible and practical to affirm that, within the framework of this definition, Aristotle conceives epieikeia –in its exact philosophical and institutional essence– to operate as a principle that fulfills its lawful purpose in the interpretation and implementation of legal norms not contra legem, but praeter legem. In other words, recourse to epieikeia does not in any way nullify the written legal norm, but serves, in favor of Law and Justice, as an interpretative mechanism during its implementation, to supplement or even correct it. In other words, either to fill its gaps or to remedy its regulatory shortcomings, whenever and wherever its intrinsic generality—according to the above—leaves such gaps or results in such defects. In this sense, as already noted, Aristotle sees epieikeia as functioning within the field of interpretation and implementation of the legal norms, as the “steward” (therapainis) of Justice, operating institutionally within the narrower scope of Distributive Justice and, where this is insufficient for the full realization of Justice, of Corrective Justice as well.

c) Finally, concerning the methods by which the principle of epieikeia is to be employed in practice, Aristotle’s analysis for the institution of Arbitration stands out for the originality of its formulation—especially considering the intellectual context of his time. He treats Arbitration as a procedure for administering Justice that functions, in a sense, in parallel with the jurisdiction of the judge, who acts as a state authority formally empowered to resolve disputes between members of the political community. Specifically, by introducing the principle of epieikeia into the institutional framework of the strictly state-based jurisdictional function, Aristotle observes, among other things: «Καὶ τὸ εἰς δίαιταν μᾶλλον ἢ εἰς δίκην βούλεσθαι ἱέναι ὁ γὰρ διαιτητὴς τὸ ἐπιεικὲς ὁρᾷ, ὁ δὲ δικαστὴς τὸν νόμον· καὶ τούτου ἕνεκα διαιτητὴς εὑρέθη, ὅπως τὸ ἐπιεικὲς ἰσχύῃ» (Rhetoric, I.13, 1374b 18–22).

c1) With this observation, Aristotle broadly acknowledges that, beyond the officially institutionalized state judicial authority, there exists—albeit with institutional and political recognition by the polis—the institution of Arbitration, as a procedure for resolving disputes between members of the social body, based on the free choice of the parties involved. He concludes that the recourse to Arbitration constitutes a form of practical application of the principle of epieikeia in the administration of Justice. That is to say, for Aristotle, Arbitration is, in effect, an institution inseparably connected to Justice. This is because, in the framework of the formally organized judicial system of the state, the judge relies primarily—if not exclusively—on law, that is, on legal rules enacted by the state. In contrast, in the context of Arbitration, the judge —as an arbitrator appointed by the two disputing parties— relies chiefly on methods of decision-making that derive from the invocation of epieikeia. Aristotle thus reaches the conclusion that the main reason, for which the institution of Arbitration was conceived, lies in resolving disputes between adversaries with the principal criterion being the use of epieikeia, aiming exclusively at serving, as fully as possible, the principles of Justice.

c2) Aristotle’s views on Arbitration—as a process of dispensing Justice through the preferential use of the principle of epieikeia, and operating, by the free choice of the disputing parties, in parallel with the judicial structure established by the state—undoubtedly constitute strong evidence of his remarkably liberal views (by the standards of his era) concerning political organization in a broader sense. This is also apparent in his overall political worldview, as developed in the Athenian Constitution. Hence, some scholars have even argued that, according to Aristotle, on the one hand, the option of choosing Arbitration is available across the entire spectrum of dispute resolution, regardless of the specific nature of the dispute. On the other hand, in cases where Arbitration is chosen, the arbitrator judges legibus solutus—that is, outside and beyond the concretely established legal norms, relying solely on his judgment in accordance with epieikeia. This further implies that the arbitrator may even reach the point of generating legal norms ex novo, potentially even contrary to those established by the formal legal order of the state.

c3) A straightforward reading of the Athenian Constitution suffices to show that such extreme views on Arbitration—and on the role of the principle of epieikeia within it—are foreign to the institutional and politically liberal mindset of Aristotle, particularly regarding the organization and functioning of the administration of Justice. First and foremost, it is evident that, for Aristotle, it is inconceivable that disputes essential to the survival of the social body—governed by the strict law enacted by the state—could be subject to Arbitration at the discretion of the disputing parties. Most notably, this applies to disputes over serious crimes in the field of Criminal Law, especially those directed against the political order itself (i.e., political crimes, and above all, crimes of impiety and treason).

c4) Accordingly, in Aristotle’s conception of Arbitration, it was primarily suited to those disputes that arose between members of the community within the sphere of what we would now—admittedly with the inevitable risk of imprecise anachronistic classification—term lato sensu Private Law. In particular, this includes disputes emerging from everyday common transactions or specific commercial dealings, most characteristically involving property and contractual transactions—naturally according to the much later established terminology, in light of the Roman Law system and the Legal Science developed upon it. Moreover—perhaps even more crucial for conceiving the core of the institution of Arbitration—it is extremely difficult to accept that Aristotle, in line with the interpretations presented above, endorsed a form of unrestricted liberalism on the part of the arbitrator in exercising this special form of jurisdiction. More precisely, such liberalism that would amount to a direct decision contra legem, therefore, a decision which would completely defy the general principles, stemming from the officially appointed legal rules suited to judge similar cases, only this time adjudicated by an official judge, bound by legal jurisdiction.

c5) Consequently, any preference Aristotle appears to express—at whatever level or in whatever degree—for Arbitration must be interpreted in conjunction with his broader conception of the nature and jurisdictional reach of the principle of epieikeia. This means that, according to Aristotle, the disputing parties may choose neither the Arbitration nor the arbitrator in order to resolve their dispute contra legem. Rather, the arbitrator, within the framework of Arbitration, enjoys a significantly broader discretionary margin, than the officially appointed judge, to issue decisions praeter legem. However, such discretion could only be exercised, in case of a legislative gap, which ought to be covered by rules, inspired and implemented by the arbitrator himself, in accordance with the principle of epieikeia. Or, when the applicable concretely established legal norms had to be supplemented or suitably corrected—again in accordance with the principle of epieikeia—in order to fully conform to the concept of Justice. More precisely, to the concept of Distributive Justice and its fundamental institutional and normative parameter: proportional Equality.

c6) In conclusion, it must be accepted that, for Aristotle, the arbitrator—within the framework of the institution of Arbitration—judged according to the choice of the disputing parties as a kind of co-actor alongside the legally appointed jurisdictional officers. The essential difference between them lies in the fact that the latter primarily adjudicate secundum legem, whereas the arbitrator had the capacity to develop his legal reasoning by utilizing all the interpretive channels available praeter legem, without, however, entirely defying or, even more so, directly violating the established Law.

II. To what extent and in what way Aristotle’s conceptions of Law and Justice clarify the limits of the intervention of Artificial Intelligence in the Legislative and Judicial Functions

In order to understand how the institutional and normative essence of Law and Justice—as previously analyzed on the basis of Aristotle’s positions—sets boundaries on the potential interventions of Artificial Intelligence in the exercise of legislative and judicial power, it is necessary, first and foremost, to briefly outline the relevant capacities of Artificial Intelligence. This necessarily presupposes a clarification of the contrast between Artificial Intelligence and Artificial Consciousness—especially because, as will be elaborated in what follows, it is the absence of Artificial Consciousness that most fully explains why Artificial Intelligence alone is incapable of handling, in a law-making or adjudicative context, the full normative depth of indeterminate legal concepts such as those embedded in Law, Justice, and epieikeia.

A. The distance between Intelligence and Consciousness

According to the currently dominant view across scientific communities—even with the emergence and still undefined prospects of quantum computing and its potentially vast capabilities—it is universally accepted that the human being is the most “intelligent” entity on the planet. This is because man is a combination of homo sapiens and homo sentiens and thus possesses not only Intelligence but also Consciousness. Even the most advanced Large Neural Networks (LNNs) possess only Artificial Intelligence, albeit of the highest order, which in some cases seems to suggest a readiness for a gigantic leap toward Artificial Consciousness. However, no such prospect appears on the horizon—even in the distant future—for the creation of Artificial Consciousness as a fully developed scientific concept itself remains, to a large extent, unexplored.

1. Into the depths of Consciousness

It is a widely acknowledged fact that the human being retains an undeniable superiority over any other entity—of any kind—on Earth, precisely because, as previously noted, man is a synthesis of homo sapiens and homo sentiens. In essence, the human being possesses, on the one hand, Intelligence, which opens a vast field of Knowledge; and, on the other, Consciousness, which—through accumulated experience gained via Knowledge and other means—leads to empathy, self-awareness, and ultimately, to self-consciousness. This is a scientific conclusion derived through the cooperation of multiple disciplines—though with the additional clarification that we still do not fully understand how the human brain functions, especially in regard to the formation and initiation of the process that activates Consciousness.

a) It is indeed generally accepted that there is a continuum between the unconscious and the conscious. In practical terms, human behavior starts out at the unconscious and reaches the stage of the conscious, at which point the process of activating Consciousness is completed. It is then that the individual initiates the corresponding process of choosing an action, an omission, or any other form of behavior.

a1) The truth is that we do not yet sufficiently understand—and it remains uncertain whether or when we will ever attain such understanding, given that scientific perspectives on the matter are themselves conflicting—the structure and origins of the unconscious, and consequently the functioning of the mechanism that enables the transition from the unconscious to the conscious. On the other hand, our adequate understanding of the structure and function of Intelligence —as a dynamic ensemble of cognitive capacities that enable human beings, among other things, to learn, to comprehend, and ultimately to formulate views, beliefs, and concepts— has permitted the emergence, and indeed the increasingly intense development, of Artificial Intelligence. By contrast, our extremely limited, as previously noted, insight into the arcana of the unconscious and conscious—and of the communication between them—renders impossible even a grosso modo approximation of the prospect of creating Artificial Consciousness. For the purposes of the present study, the only reliable basis we can rely on is a working description of Consciousness. I adopt here the most recent and relevant definition offered by Christof Koch (The Feeling of Life Itself: Why Consciousness Is Widespread but Can’t Be Computed, The MIT Press, 2019, p. 1): “Consciousness is experience … lived reality. The feeling of life itself”. More elaborately, consciousness is the property and capacity of the nervous system to produce a unified set of feelings and beliefs that lead to self-awareness—of both our own selves and of our surroundings. Furthermore, the only point that is scientifically substantiated with sufficient certainty is that the operation of Consciousness depends crucially on the involvement of a vast number of neurons present in the human body—and in particular in the central nervous system.

a2) It is thus entirely utopian to make any attempt to resolve this Delian problem—the quintessence of the unconscious and the conscious—through the advancement of Artificial Intelligence or the future programming of some exceptionally “intelligent” LNN (Large Neural Network). Such programming would presuppose a sufficient comprehensive understanding of the parameters of the problem itself; in other words, it would require adequate knowledge of the unconscious–conscious mechanism, which, as explained above, does not exist.

b) Consequently, no LNN —including the so-called Large Language Models (LLMs)— possesses Consciousness, not even in an embryonic form. What actually occurs is that these systems are so advanced that they are capable of accumulating, through appropriate and continually evolving programming, vast amounts of Knowledge. And through this accumulated knowledge, they can—again via proper programming and reprogramming—assist humans in solving extremely complex problems, including those involving the development of effective and efficient methods of decision-making. These methods are often highly significant for the future and prospects of entire nations—and indeed, of the planet as a whole. It is notable that, in this domain of the application of Artificial Intelligence and of the accumulation of knowledge, LNNs and LLMs are tentatively, in proportion to the rapid evolution of their programming, able to surpass even humans—however, only in certain, specific domains.

b1) In particular, with regard to what was briefly mentioned earlier concerning the potential and contribution of the Quantum Computer, the following clarifications must be added: The Quantum Computer, which is still essentially in an embryonic stage with respect to its future performance across a vast array of technological fields, is a product of Quantum Technology and its possibilities go far beyond those of Artificial Intelligence. As for their technological differentiation—bearing in mind the important caveat that researchers in Quantum Technology generally do not, at least for the time being, integrate Artificial Intelligence into their methods—it can broadly be said that Artificial Intelligence makes the “machine” more intelligent, whereas Quantum Technology, with the most tangible example being the Quantum Computer itself, makes it more rapid. Furthermore, in conclusion, the speed of the Quantum Computer may enhance to an unimaginable degree the “intelligence” of AI systems, performing rapidly calculations which, until now, required incomparably more time. Thus, it secures for AI systems incomparably greater capabilities of extremely fast programming and reprogramming. This could greatly facilitate the formulation of predictions or the discovery of entirely new paths across the entire spectrum of scientific research. In this regard, the horizon of gradual development appears increasingly expanded, both in the near and —most importantly— in the distant future.  What becomes quite certain is that the Quantum Computer introduces wholly new, and as yet undefined, possibilities for the application of Technology, particularly in the field of Game Theory and, further, in its multidimensional applications.

b2) However, despite the fact that, as explained above, no one is justified to underestimate the almost awe-inspiring developments surrounding the use of the Quantum Computer, there is absolutely nothing—not even under the broadest spectrum of hypotheses and speculative projections—that allows for even the slightest suggestion that the Quantum Computer and its applications will make possible the transition of Technology from the stage of Artificial Intelligence to that of Artificial Consciousness. What is certain is that the Quantum Computer may push Artificial Intelligence far beyond its current limits—but not so far as to cross the Rubicon towards Artificial Consciousness. All indications suggest that neither Quantum Technology nor the Quantum Computer is, intrinsically, capable of fulfilling such a task. Accordingly, here too the interdisciplinary maxim applies: impossibilium nulla obligatio est.

b3) As for the extreme limits of the capacities of Artificial Intelligence through the combination of all currently available means one more observation must be added: As noted by the authors of The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies (Erik Brynjolfsson and Andrew McAfee; Greek trans. Giorgos Nathanael, Kritiki, Athens 2016), despite the significant progress of Artificial Intelligence in recent years, researchers working in the field have still not managed to refute the famous Moravec Paradox, first articulated in the 1980s by Hans Moravec, a pioneer in the field of Robotics and Professor at Carnegie Mellon University. According to this paradox, “it is relatively easy to make computers perform well on intelligence tests or in chess competitions, but difficult to impossible to equip them with the perceptual and motor skills of a one-year-old child.” (See especially Hans Moravec, Mind Children, Harvard University Press, 1990; and Mere Machine to Transcendent Mind, Oxford University Press, 2000).

 2. Under the Constellation of Erosive Risk

For the sake of completeness in the present analysis, it is important to emphasize that as necessary as research for the enhancement of Artificial Intelligence capabilities may be in our time, equally necessary—if not more so—is research into the potential risks entailed in the unrestrained use of AI tools, particularly when such use leads to the making of critical decisions which affect not only the future of individual human beings but of Humanity as a whole. An especially telling example is the AI program AlphaZero, developed through a Large Neural Network (LNN) to play high-level chess.

a) The AlphaZero program represents a milestone for AI, wherein the programming of an LNN with chess games allowed it to generate new games and strategies never before conceived by humans. Moreover, it enabled the system to play these games in such a manner as to consistently outmaneuver its human chess-player opponent—always catching them off guard and always winning. This element of surprise stems largely from the unpredictability of the LNN’s moves (e.g., the “sacrifice” of a queen at a moment a human chess-player, however experienced, would never contemplate and make such a move), which completely dismantles the human opponent’s strategy—leading, at best, to a draw, but more likely to defeat.

b) The AlphaZero example also leads us into a deeper and more pressing concern for the dangers which arise when AI is deployed at such a high level via LNNs. These systems are programmed—and, due to the highly advanced nature of contemporary AI, can be reprogrammed almost indefinitely—to solve the problems they have been assigned with, particularly the problems involving decision-making. This implies that AI algorithms, as their programming is being advanced in accordance with technological evolution, can solve increasingly complex problems. As their underlying logic improves, these algorithms generate newer and functionally more sophisticated forms, in the field of Artificial Intelligence. In effect, the LNN is updated and enhanced so as to seek and find a solution—by any means available.

c) However, as previously noted, apart from Artificial Intelligence, an LNN does not possess nor are we able to equip it with Artificial Consciousness, in any way or form. Consequently, it lacks the autonomous capacity to formulate or follow moral principles concerning whether the solution it proposes may be harmful or even catastrophic to humans, who exist within a social whole. This necessarily means that an LNN can, at most, recognize only those ethical norms and those ethical constraints, which are programmed by the human agents who design its algorithms. An LLN is, in this sense, externally determined and cannot possess moral constraints, of its own accord. Here lies the critical and profoundly troubling a priori reversal: the trajectory of Artificial Intelligence, via LNNs, continues unabated, even though we have yet not formulated even the most basic objective norms for their deployment, so that they do not jeopardize the future of humanity and, ultimately, of our own Civilization.

B. Some dangerous illusions in the field of Legal Science

What has been outlined above illustrates just how far removed from reality are those who foresee in the field of Legal Science a mode of utilizing contemporary LNNs and LLMs in such a way as to allow them to substitute, even in certain cases, the institutional members of the Legislative, Executive, and Judicial powers—either to produce legal norms or to interpret and implement them. This concern applies especially to the officials responsible for the exercise of adjudicative authority in general. The truth is that AI can be of assistance to legal reasoning only in a supplementary way, primarily in cases where the interpretation and implementation of the relevant legal rules require the resolution of complex technical problems, e.g., in fields such as Economics, Physics and Mathematics, or during evidentiary procedures in the exercise of judicial functions, particularly through expert analysis. In such contexts, AI may assist officially authorized legal practitioners to interpret and apply the law. This limitation arises significantly, among other factors, because Artificial Intelligence lacks even the most rudimentary characteristics of Consciousness. As a result, AI is structurally and by definition incapable of overcoming the technical and legal barriers involved in either: (i) the effective and lege artis construction of indeterminate legal concepts in the formation of legal norms, or (ii) the concrete specification of such concepts in accordance with their normative function during the administration of justice. Returning, therefore, to Aristotle’s previously outlined views on Law, Justice, and indeterminate legal concepts, we may draw the following conclusions:

1.The distinctive nature of Law-Making and Judicial Functions

First and foremost, the so-called automatic lawmaking —especially the reliance on supposedly automatic legal services and judicial decisions— is not only unfounded but profoundly dangerous for the very essence of Law and Justice as they are meant to function, in accordance with the Aristotelian framework already discussed.

a) This is particularly true because the Rule of Law—within the framework of the modern Rule of Law state and the Principle of Legality, and therefore as an essential normative component of Law and Justice—is, by nature and by definition, instituted for the regulation of ever-evolving and interdependent human behaviors within a given social body. This means that its establishment, interpretation, and implementation necessarily and intrinsically entail a complex and multifaceted legal and factual evaluation, particularly due to the quasi-admissible interference of a multitude of indeterminate concepts, both legal and evaluative. Such an evaluation, under the conditions described above (i.e., decision-making, which respects the legal formality and chooses the adequate interpretation of the law) almost always requires the activation of the transition mechanism from the unconscious to the conscious. Further, it presupposes, in any case, the combination of Intelligence and Consciousness. Indeed, given the previously discussed nature of the democratically established and officially authorized Rule of Law, it is virtually inconceivable to imagine situations in which its establishment, interpretation, or implementation could occur solely through the function of Intelligence, therefore in the absence of Consciousness. The normative constitution of the Rule of Law itself —through the necessary interweaving of the factual and the legal which determines the entire process from establishment to implementation— and its inclusion in the normative category of indeterminate legal concepts therein, render such a possibility not merely marginal and negligible, but altogether unthinkable.

b) Under these conditions, only the human being —acting, in each case, as a competent official endowed with both Consciousness and Intelligence— is capable, on the one hand, of creating Law through appropriate and democratically legitimate legal norms and, on the other, of administering Justice in the full sense of the term and in whatever form. By contrast, LNNs and LLMs, which —as previously demonstrated— lack Artificial Consciousness, are incapable of fulfilling such a mission in its entirety, at least in accordance with the normative purpose of the Rule of Law and Justice in a democratically organized and governed society. Consequently, Artificial Intelligence can operate within the domain of Legal Science only in a supplementary way and within well-defined limits —never in a primary or autonomous role—, whether this involves the establishment of legal norms or their interpretation and implementation. This is primarily due to the fact that AI is structurally unable to overcome the pervasive challenge posed by the need to process a multiplicity of indeterminate, legal or evaluative concepts, which must be specified in concreto (case by case), as already explained.

c) This, in turn, entails —by strict legal logic— that legal norms and judicial rulings or decisions —including those rendered by arbitral tribunals, insofar as they too are called upon to specify indeterminate legal concepts within the scope of their special jurisdiction, particularly on the basis of the principle of Equity, which, as previously noted, is itself a normatively indeterminate legal concept— when generated automatically by Artificial Intelligence, negate the very essence of Justice. Especially when such automation would mathematically result in the exaltation of a generalized, leveling regulatory treatment of social relations, entirely at odds with the Principle of Equality in its proportional sense and, furthermore, of the inherent value of the Human Being and the free development of their individuality.

c1) Furthermore, LNNs and LLMs, for the same reasons, are incapable—even under the most advanced and complete programming— to establish legal norms that could be integrated into a given legal order (stricto or lato sensu) in a manner that fully satisfies the normative requirements of its hierarchical structure. For in any democratically organized State, where the principle of Separation of Powers is properly upheld —along with the corresponding principles of the Rule of Law and Legitimacy— the structure of the Legal Order is based on norms arranged in a hierarchy of normative force and scope.

c2) However, as follows from the foregoing, and considering the current lack of Artificial Consciousness, even the most advanced LNNs and LLMs do not possess the complex cognitive capacities —understood in the broadest possible sense— which, through the transition from the unconscious to the conscious, would enable the documented and comprehensive assessment of whether a legal norm they may generate conforms, for example, to the Constitution in the context of a national legal system, or to International and European Law, should one be operating within the normative framework of the international community or the European Union. This is because no one can ignore the current reality, in which every national legal order coexists and cooperates necessarily with the international legal order —and, in the case of EU Member States, with the European legal order— with all the resulting consequences and demands for exceedingly complex legal reasoning concerning the establishment, interpretation, and implementation of specific legal norms.

2.The value of Kant’s philosophical reflection

Here, once again, the great contemporary significance of I. Kant’s philosophical reflection is revealed—this time in yet another scientific domain. Only the human being, as a creature endowed not only with Intelligence but also with Consciousness, is capable of establishing rational rules of conduct, which apply both to himself and to his fellow human beings. These are rules which conform to the model of the “categorical imperative” and its five forms, as set out by Kant. That is, rules of conduct which are both formed through the free will of their creator on the basis of Morality and are applied in practice in accordance with the demands of Morality. In other words, rules of behavior built upon ethical constraints — constraints which the means of Artificial Intelligence could, hypothetically, generate only if a kind of technological “transfusion” of Artificial Consciousness had been achieved, as described above.

a) Consequently—as well as for the additional reasons set out above—LNNs and LLMs are incapable of autonomously producing complete legal rules in the form of leges perfectae, and not merely in the deficient forms of leges imperfectae or leges minus quam perfectae. Nor can they interpret and apply such rules mechanically, due to their lack of Artificial Consciousness. There are many examples to reinforce this, both in the domain of General Legal Theory and also as regards the interpretation and implementation of norms within Private and Public Law. Even at this initial and broad level of inquiry, we must accept the general conclusion that LNNs cannot reach the level of conceptual development required, for example, to formulate general principles of law as sources of the legal system derived from the legal norms in every Legal Order as a whole, nor to appropriately apply the principle of Equity in the interpretation and implementation of law during the administration of Justice.

b) This incapacity becomes even more evident in areas such as:

b1) Constitutional Law, whose normative authority is supreme due to its foundational status at both the base and apex of the Legal Order. From this alone, one can understand the severely limited contribution of LNNs and LLMs in deriving general constitutional principles and in interpreting open-ended constitutional clauses —such as the clauses for the respect and protection of the dignity of Man (Article 2 par. 1 of the Constitution), the free development of personality (Article 5 par. 1), and the principle of proportionality (Article 25 par. 1.d).

b2) European Law, the typical provisions of which also normatively prevail over those of ordinary legislation. Here too is made evident, for the reasons already discussed, the limited contribution of LNNs and LLMs in deriving general principles that have equal normative force with the rules of both primary and secondary European Law.

b3) International Law, whose provisions also take precedence over those of ordinary legislation, though under different conditions. This is particularly evident in the context of the virtually nonexistent capacity of LNNs and LLMs to derive generally accepted rules of International Law, as defined by Article 28 par. 1 of the Constitution. Such a derivation presupposes, among other things, an extremely nuanced interpretation of the existing norms of International Law — especially from the standpoint of its teleological, systematic, and historical dimensions.

c) Furthermore, LNNs and LLMs do not possess the ability, for example:

c1) To formulate from the ground up and interpret legal acts of Private law in a manner that adequately ensures their consistency with the principles of good faith and moral behavior; to properly apply and integrate the precepts of shared experience or common knowledge; to recognize commercial customs; or identify what constitutes the “due diligence in one’s own affairs” (ἐν τοῖς ἰδίοις ἐπιμέλεια).

c2) As regards the interpretation and implementation of Public Law, to precisely grasp the meaning of concepts such as democratic legitimacy, proportionality —especially in relation to the officially constituted restrictions on rights— good governance, and the protection of the citizen’s trust in government. Or to exercise discretion in full, since discretion intrinsically and by definition involves the weighing of indeterminate evaluative concepts, therefore, a in concreto (case-specific) assessment and balancing to arrive at a decision. The use of the means of Artificial Intelligence to assist state officials exercise their legally delegated discretion lege artis is a wholly different matter, particularly in scientific and specialized technical domains. In such cases, the AI contribution is supportive and not primary, since the final judgment as to the exercise of discretion belongs to the state officials in question.

d) The same holds true —a fortiori— in the Public Law domain in respect with the comprehension and evaluation of one of the most important indeterminate legal concepts regarding state action in general, i.e. Public Interest. This becomes all the more evident as state action increasingly extends into highly specialized technical domains of intervention.

d1) It is particularly telling that LNNs and LLMs are intrinsically incapable of determining —at least not as adequately as required by the principles of the Rule of Law and Legitimacy— whether in a specific case what state officials refer to as Public Interest is in fact a disguised form of purely Fiscal Interest. Accordingly, it becomes readily apparent just how incapable LNNs and LLMs are to counter effectively one of the great current menaces, which undermines both the Rule of Law and the principle of Legitimacy, namely the “monetary sovereignty” over the “institutional sovereignty”.

d2) This is because they cannot distinguish, as already noted, between genuine, i.e. authentic, Public Interest, on the one hand, and purely Fiscal Interest, on the other, which aims to advance merely the financial goals of the state. And only in a secondary phase —through the legislative process— might that Fiscal Interest eventually rise to the level of genuine Public Interest, e.g., via the allocation of budgetary resources explicitly designated for a public purpose.

C. Conclusions Based on the Normative Structure of Judicial Reasoning

The preceding observations clearly demonstrate just how marginal the prospect is for any meaningful intervention by Artificial Neural Networks (ANNs) and Large Language Models (LLMs) in what is perhaps the most critical domain of Legal Science and its practical applications — namely, the administration of Justice and, specifically, the judicial decisions being issued by the official judicial authorities under jurisdiction in each case. For Legal Science, in accordance with its fundamental purpose, must not be conceived solely as an exercise of useful, however not independently functional, theoretical constructions, related to the normative particularity of each Legal Order. This is because, in general terms and in accordance with its institutional essence, Law is enacted to be applied in line with the imperatives of Right and Justice. The theoretical elaboration of its normative dimensions is not an end in itself but rather, in a sense, a “steward” to the application of the Law in practice, as dictated by its origin and the purpose of its enactment—hence the primary importance of its teleological interpretation. Against this background, analysis in this direction becomes all the more significant—reaching a level of heightened urgency—given that for over thirty years now, primarily in the United States, practices involving the institutionalized use of ANNs and LLMs have been adopted and widely promoted, with the goal of fully replacing judicial processes in corresponding fields. This includes the almost entirely automated issuance of purported judicial decisions, achieved through appropriate programming.

1.The Inherent Limits of Judicial Reasoning in Relation to Its Programming via Algorithmic Predictions

Referring back to the aforementioned discussion concerning the normative nature of various legal concepts and data, within the context of interpreting and applying the operative legal norms in practice, it becomes—indeed, indisputably—evident why traditional judicial reasoning, in the administration of Justice by the judicial authorities vested with in concreto jurisdiction, cannot conceivably be carried out exclusively by ANNs and LLMs, even when programmed—at the most advanced levels—using methods of algorithmic prediction. Beyond the normative character of legal concepts—and particularly of indeterminate legal concepts—and legal data, and as a direct consequence thereof, the very structure and function of judicial reasoning, in all its dimensions, makes evident its inherent limitations when it comes to programming it through algorithmic predictions. Specifically:

a) First, from an institutional standpoint and for judicial reasoning to be articulated lege artis, it is essentially impossible to achieve a fully accurate selection of the applicable legal rule—the so-called major premise of judicial reasoning—through algorithmic predictions.

a1) As a preliminary clarification, the major premise of judicial reasoning consists in the selection of the applicable legal rule—or a combination thereof—for resolving a dispute brought before a judicial authority with jurisdiction. In this respect, the major premise constitutes, according to legal logic, the sedes materiae of judicial reasoning, with the added consideration that the selection of the applicable normative framework is inextricably linked, on the one hand, to the ongoing flow of legislative production by the Legislative Power and, on the other, to the stream of factual circumstances that gave rise to the legal dispute in question. These circumstances must, for the dispute to be resolved, be subsumed under the chosen normative framework.

a2) Contemporary reality emphatically demonstrates the relativity and, therefore, the uncertainty that characterizes the rapid transformation—primarily driven by economic and technological developments—both of the socio-economic reality, which plays the role of the regulatory “infrastructure” for the applicable legal norm, and of the legal norms themselves, which are established as the institutional “superstructure” of that infrastructure and—quasi-deterministically—are continuously supplemented, amended, or even replaced so as to effectively fulfill their normative function. This striking and increasingly intense fluidity, which affects both the “infrastructure” of socio-economic reality as well as the “superstructure” of the legal norm(s), strongly substantiates the previously stated conclusion: that it is no longer institutionally feasible—given the nature of judicial reasoning and its direct connection with the imperatives of Justice—to program its major premise through algorithmic predictions. Moreover, this limitation cannot be adequately overcome, even through the most specialized categorization of the disputes under adjudication, since the above-noted fluidity regarding the foundations of the major premise of judicial reasoning arises from their very nature, not from the degree to which they may be concretized. For this reason, as well, it is impossible—under any interpretation—to conceive of a form of judicial reasoning with a quasi-absolutely binding conclusion that could render feasible the exclusive programming of its major premise via algorithmic predictions.

b) Second, the considerations already elaborated concerning the programming of the major premise of judicial reasoning suffice to demonstrate that ANNs and LLMs are not, from an institutional standpoint, capable—nor should they be permitted—to substitute the judicial authorities constitutionally entrusted with the administration of Justice.

b1) This conclusion is reinforced a fortiori by the normative character of the minor premise of judicial reasoning, all the more so because it is inextricably linked to the major premise. The minor premise of judicial reasoning consists, fundamentally, in the interpretation—in all its aspects, beginning with the grammatical and, notably, the teleological—of the legal rule(s) forming the major premise, and, on that basis, in the lato sensu legal classification of the factual circumstances that emerge during the evidentiary process, in accordance with the relevant procedural rules.

b2) Under these conditions, the correlation between the major and minor premises of judicial reasoning further clarifies—and indeed reinforces—how and why even the minor premise is not amenable to full and comprehensive programming through algorithmic predictions. As has already been noted, such programming may prove extremely useful only with respect to the evidentiary process within judicial reasoning during the resolution of a dispute—particularly when the nature of the dispute involves the resolution of issues which require recourse to the empirical data of the Natural Sciences, and chiefly of Mathematics and Physics.

c) Third, the infeasibility of fully programming, via algorithmic predictions, both the major and the minor premises of judicial reasoning inevitably extends to its third component—namely, the conclusion. Thus, the accuracy of the initial claim becomes even clearer: that ANNs and LLMs cannot—at least for the foreseeable future, given that we have not reached, and likely cannot foresee whether we shall ever reach, the stage of creating Artificial Consciousness—become authentic judges of the present or the future.

d) At this point, it must be added that the analysis of the impossibility of programming judicial reasoning via algorithmic predictions applies in full across all levels of jurisdiction, including judicial review by abrogation. It must also be emphasized that the digitization of court operations, aimed at facilitating the administration of Justice, is a wholly distinct issue from the formulation of judicial reasoning itself. In truth, Artificial Intelligence can be of significant assistance to judicial authorities and litigants alike — particularly in providing access to current legislation and existing case law, facilitating communication between litigants and the courts (e.g., electronic filing of documents), enabling access to case files, and supporting the overall organization of court calendars and the preparation of dockets.

2.Legal antibodies against the attempted robotization of Justice

As noted at the beginning of this section, it is primarily in the United States that we observe the phenomenon of an expanding institutional entrenchment—albeit limited to certain domains of judicial decision-making—of increasingly frequent judicial decisions being issued through the full programming of judicial reasoning via pre-constructed, extensive algorithmic predictions. The foregoing argumentation is sufficient to demonstrate both the untenability and the dangers of such practices with respect to the administration of Justice. What follows, therefore, is necessary, primarily inasmuch as it reinforces the positions advocating the urgent need for developing the necessary antibodies to prevent a spread, profoundly corrosive to the administration of Justice, which would extend far beyond the USA, e.g. to Europe and the Legal Order of the Member States of the European Union, regarding the automation of the organization of the judiciary. Such a development would be corrosive also to the Rule of Law and the effective exercise of the Fundamental Right to judicial protection.

a) By way of introduction, it should be noted that the tendency in the U.S. to adopt methods in order to issue judicial decisions through fully programmed algorithmic predictions may be partially justified by the ongoing importance, within Anglo-Saxon Law, of the legal doctrine of “precedent”.

a1) The judicial institution of the “precedent” allows, in certain cases, the competent adjudicating authority to base its decision almost exclusively on previous judicial rulings, primarily those of higher or supreme courts. It is precisely this structural feature of precedent that highlights the key difference between the Anglo-Saxon Law and, for example, the normative structure of the Legal Orders of Continental European States. For under such conditions, the “precedent” constitutes concrete proof that the Judiciary holds a distinctly law-creating function. That is, it entails a direct acknowledgment that Courts possess the actual, jurisprudential power to shape entirely new legal norms.

a2) This, as already clarified, does not apply at all in the Legal Order of the Continental European States. Given these characteristics, the “precedent” in the context of the judiciary—for instance, in the U.S.—appears, prima facie, to facilitate the programming of judicial decisions via algorithmic predictions. This facilitation is deemed effective because, allegedly, the element of contingency in judicial reasoning seems to be diminished, if not eliminated altogether, since the “precedent” leaves little room for the court with in concreto jurisdiction to deviate from its legal and factual determinations.

b) However, this understanding of the doctrine of the “precedent” is, without doubt, superficial and, further, disingenuous—aimed at indirectly or even directly legitimizing the use of methods for automating judicial decision-making under the conditions already clarified.

b1) This is primarily because, aside from the clear fallacies of an entirely static conception of it, the “precedent” can no longer, even within Anglo-Saxon Law, fulfill the role it held in the more distant past. Indeed, the phenomenon—analyzed extensively earlier—of the erosion of the normative force of legal norms, due on the one hand to the rapid transformation of their socio-economic “infrastructure” and on the other to the inflationary production of new legal rules, directly impacts the field of Anglo-Saxon Law as well. Consequently, this complex normative foundation of the “precedent” —which in the past had guaranteed unabated its procedural effectiveness—has been largely stripped of the solid institutional basis on which it had been established within Anglo-Saxon Law. Today, the regulatory framework comprising the major premise of judicial reasoning no longer possesses the stability that once characterized it, and likewise, with respect to the minor premise, the socio-economic conditions no longer lend themselves to the level of predictability that the “precedent” requires.

b2) Under these circumstances, those who invoke the once-prevailing concept of the “precedent” to justify the adoption of automated judicial decision-making methods appear either to ignore—or deliberately overlook—the substantial contemporary transformations it has undergone, thereby rendering the purported authority of fully programmed justice through algorithmic predictions legally dubious. In conclusion, even within the Legal Order of the United States, the potential recourse to the judicial services of ANNs and LLMs is, in essence, merely complementary—assuming, of course, that the above Legal Order continues to operate under the principles of a democratic Rule of Law and in accordance with the imperatives of Law and Justice.

 3.Artificial Intelligence and its technological applications as “assistants in the fulfilment” of Justice within the exercise of Judicial Authority

In conclusion—and summarizing the preceding discussion on the inherent limits imposed by the normative structure of judicial reasoning on any form of its programming through algorithmic predictions—the following must also be accepted: Artificial Intelligence, along with all its technological applications, may be deemed permissible only insofar as they serve as assistants in the fulfilment of Justice within the framework of the judicial task as exercised by its institutional functionaries. This acceptance must be understood in light of the fundamental principles of the Rule of Law and the sanctioning mechanisms which effectively safeguard the institutional purpose of Justice in accordance with the core mandates of modern Representative Democracy. In what follows I shall give an outline of the several parameters along which Artificial Intelligence and its technological applications may evolve as “assistants in the fulfilment” of Justice within the domain of judicial authority, in a manner consistent with the nature of adjudication.

a) First and foremost, Artificial Intelligence and its technological applications can beneficially contribute to the organization and operation of the judicial system as a whole, enhancing its effectiveness with respect to the handling of its interna corporis —i.e., its day-to-day internal functions— in terms of transparent acceleration and improved secure communication between the judicial system in concreto (on a case-by-case basis) and the litigants. The guiding principles of such a contribution by Artificial Intelligence and its technological applications have been highlighted, for instance, by the Consultative Council of European Judges (CCJE) in its Opinion No. 14 (Strasbourg, 7–9 November 2011), which clarifies, among others, the following:

a1) First, Artificial Intelligence and its technological applications provide judicial officers themselves with appropriate tools to facilitate the administrative functions of courts so that such functions may proceed swiftly and effectively —particularly in coping with the high volume of cases— without compromising the integrity in the administration of Justice as determined by the fundamental principles of the Rule of Law and Legitimacy.

a2) Second, Artificial Intelligence and its technological applications are capable of making more productive, in quantity and, especially, in quality, the relationships between judicial officers—and thus the courts—and other public services which require the relevant information and data in order to exercise their functions. This is especially true with regard to certain public services which operate, in support of the judiciary, in the area of out-of-court dispute resolution among members of a certain social body.

b) Secondly and consequently, Artificial Intelligence and its technological applications can play a decisive role in upgrading the overall policy of “Justice governance,” both in terms of its internal functions and in relation to its interactions with third parties involved in the administration of Justice. A pioneering contribution in this regard was made by the European Commission for the Efficiency of Justice (CEPEJ), through the set of guiding recommendations issued at its 28th Plenary Meeting, held on 7 December 2016, under the title “Guidelines on How to Drive Change Towards Cyberjustice”. These guidelines address the following key areas:

b1) First, the area of access to Justice. This includes, on the one hand, providing information about the rights of the litigants in particular, as well as access to data concerning existing case law. On the other hand, it provides access to specific procedures related to out-of-court dispute resolution, most notably mediation.

b2) Second, the area of lato sensu cooperation between the courts and all sorts of Justice “collaborators“, primarily lawyers. In this regard, it is important to highlight the role of technology in facilitating those who collaborate with the judiciary in identifying and submitting the appropriate legal remedies and means, as well as in substantiating the corresponding pleadings by enabling access to relevant jurisprudential data.

b3) Third, the area of court administration in terms of the day-to-day execution of their judicial mission. This includes, for instance, matters such as the filing of pleadings, the tracking of the stage in a procedure, the creation and continuous updating of court hearing dockets, and the facilitation of judicial conferences. In this context, the use of telematics can offer tremendous support to judges, helping ensure the timely adjudication of cases.

b4) Fourth, the area of assisting judicial officers primarily in their ongoing professional training and their jurisprudential awareness—both with respect to domestic case law (intra muros) and that of foreign courts beyond national borders. For our own judiciary system in particular, such assistance becomes especially critical when it comes to jurisprudential data from the Courts of the European Union—and most notably the European Union Court of Justice—as well as the European Court of Human Rights, within the framework of applying the provisions of the European Convention on Human Rights.

Epilogue

The foregoing analysis thus serves to illustrate both how and why Aristotle’s thought on Law and Justice —and furthermore, on Equity— contributes to identifying and defining the real limits of Artificial Intelligence in relation to lawmaking and the administration of Justice. Specifically, the arbitrary or even improper use of the methods of algorithmic prediction in legislation and in the implementation of legal processes with regard to judicial reasoning already reveals its destructive effects—particularly in the form of monstrosities within the field of adjudication—in domains which are critical for safeguarding the value of the Human Being and for defending the free development of the personality of an individual.

A. This observation is by no means exaggerated when one considers, for instance, that in the United States the extensive use of the tools of Artificial Intelligence in the issuance of judicial decisions is not limited to areas of minor importance within socio-economic life—and even there, as explained, it fails to prove effective—but extends far beyond such domains. It has gained ground even in the adjudication of disputes which affect critical aspects of Fundamental Human Rights, without showing any signs of responsible self-restraint. This conclusion is supported by the fact that, in specific cases in the United States, judicial decisions have been issued through programming based almost exclusively on algorithmic predictions, even in the highly sensitive domain of Criminal Law, which is of paramount importance for the protection of Fundamental Human Rights. The “jurisprudence” that has emerged in this area thus far reveals decisions which clearly exhibit a tendency towards unthinkable racism, particularly against citizens of various ethnicities. This occurs because such decisions are based —by necessity, given the nature of algorithmic predictions— on purely subjective or blatantly arbitrary criteria of classification within the so-called “risk scales.” The automation of such scales inevitably leads to equally subjective or blatantly arbitrary conclusions —e.g., in assessing the objective and especially the subjective aspects of a crime before the issue of a sentence, and in evaluating possible recidivism (see, indicatively, L. Kanellou, Applications of Artificial Intelligence in Law and Judicial Practice, Legal Library Publishing, Athens 2021, esp. pp. 171 ff., with further extensive references).

B. Thus, it becomes more than evident that, in the name of a supposed acceleration of judicial proceedings through the unregulated use of technology—and indeed, as can be readily inferred from a deeper analysis of the above-mentioned practices, without any substantial understanding or awareness of the limits of Artificial Intelligence and its relation to the essence of consciousness—the Judiciary is being dangerously distanced from its foundational roots. And primarily from those roots that bind it to the broader institutional framework of Representative Democracy as a guarantor of Freedom—in other words, as a guarantor of the unimpeded exercise of Fundamental Human Rights. This outcome constitutes yet another reason why we must acknowledge the institutionally and politically invaluable classical character of Aristotle’s thought, particularly in relation to the essence of critical and perennial parameters of Representative Democracy—namely, Law, Justice, and their normative “companion”, Equity. At this point, it must be emphatically stressed that we must never lose sight of the fact that the operation of the Judiciary —and, moreover, the exercise of jurisdiction by its officers— in accordance with the essential principles of Law and Justice, is a crucial factor for safeguarding social cohesion and for the effective strengthening of democratic institutions.