Horst Eidenmüller1Horst Eidenmüller is a Statutory Professor for Commercial Law at the University of Oxford and a Professorial Fellow of St. Hugh’s College, Oxford. He is also a Member of the Berlin-Brandenburg Academy of Sciences and Humanities (BBAW) and a Research Member of the European Corporate Governance Institute (ECGI).
* * *
A part of the series, Personalized Law.
Big data and advances in Artificial Intelligence (AI) have made it possible to personalize legal rules. In this Essay, I investigate the question of whether laws should be personalized. Professors Omri Ben-Shahar and Ariel Porat argue that personalized law could be a “precision tool” to achieve whatever goal the lawmaker wants to achieve. This argument is not convincing. The most “natural” fit and best normative justification for a personalized law program is welfarism, or utility maximization. This is because personalized law and welfarism are both based on normative individualism. But welfarism is a highly problematic social philosophy. Against this background, it becomes clear why personalized law should only have a limited role to play in lawmaking. The focus of state action should not be the design and running of a personalized law program. Rather, it should be on controlling “wild” personalization by powerful private actors.
* * *
In 1980, the Hofstra Law Review ran a symposium on “Efficiency as a Legal Concern.” Some scholars attempted to justify the pursuit of efficiency in lawmaking and adjudication. Others objected, some of them strongly. Ronald Dworkin investigated the philosophical foundations of the economic analysis of law in a piece entitled “Why Efficiency?”. He argued that there was no good answer to this question. As we all know, this did not prevent “Law and Economics” from becoming a major force in legal academia and the practice of the law—in the United States and beyond.
At the time of the symposium, there was no Internet. The first personal computers had just made it to the market. Artificial Intelligence (AI) was a specialist subject which had achieved little success with “expert systems.” Big data and the Machine Learning (ML) revolution in AI were science fiction compared to today.
That has all changed fundamentally in the last decade, and it continues to change at an accelerating pace. Technological development might be on an exponential trajectory which is becoming not only mind-boggling but also unmanageable—at least for most of us.
This is the background to Professors Omri Ben-Shahar and Ariel Porat’s ambitious and pathbreaking book, Personalized Law: Different Rules for Different People. Big data and advances in AI, especially ML, have made it possible to do something which was unthinkable until recently, namely to tailor legal rules to the individual needs and preferences of different humans. Speed limits, standards of care in tort law, consumer protections, age restrictions for risky activity, and even voting rights could be personalized, according to the authors.
But should they be personalized? Ben-Shahar and Porat are undecided: They start out by suggesting that personalized law “will” (emphasis added) change the status quo and conclude that they “wanted to persuade readers that personalized law is a worthy conversation” (emphasis added). There is a significant difference between advocating implementation of a revolutionary program of law reform and engaging in a critical discussion about the merits and demerits of such a program.
This ambiguity permeates the whole book. And it stands in stark contrast to how the proponents of “Efficiency as a Legal Concern” pitched their case in the 1980 symposium mentioned above. The advocates of the economic analysis of law, especially Richard Posner, were unequivocal about the benefits and desirability of efficiency as a guiding normative standard for legal scholarship, lawmaking, and adjudication. By contrast, Ben-Shahar and Porat show us a “brave new world” which they themselves, it appears, are not fully committed to bringing about and inhabiting.
This is not surprising. “Uniformity of the law is so universally accepted that it is little remarked,” as the authors observe. The United Nations’ “Universal Declaration of Human Rights” stipulates in Article 1 that “[a]ll human beings are born free and equal in dignity and rights” (emphasis added). It elaborates in Article 2 that “[e]veryone is entitled to all the rights and freedoms set forth in this Declaration, without distinction of any kind . . .” (emphasis added). Personalized law flies in the face of these and similar stipulations which reflect fundamental values central to democracies around the world.
Ben-Shahar and Porat acknowledge that the personalized law project “is a massive undertaking that would radically change law and raise a plethora of challenges. Why do it?” they correctly ask. Their answer is simple: “personalized law could better fulfill the goals of any system of rules” (emphasis added). They call this the “precision benefit” of personalized law. This idea is developed in a short, twenty-one page chapter of the book. (The core of the chapter on the benefits of personalization and personalized law comprises a mere eight pages). By contrast, the remaining 220 pages deal with conceptual issues, examples of “personalized law in action,” constitutional problems, and difficulties of implementation.
This is a striking asymmetry. If a few pages suffice to justify a revolution of our legal systems, then either the revolutionary idea is just utterly brilliant and evidently correct. Or, conversely, something is really problematic about it, and the best strategy appears to be to avoid going into too much substantive detail and opening yourself up to attacks.
In this Essay, I ask the question: Why personalized law? I will argue that this is a crucially important question and that Ben-Shahar and Porat’s answer to it is not convincing. The “precision benefit” of personalized law is a distraction. The authors’ argument ducks the crucial normative issue of what inspires the personalized law project. The best justification for personalized law is utilitarian: maximizing social welfare. Hence, the personalized law project brings us back to “Efficiency as a Legal Concern.” But it would be wrong to conclude that it only offers us old wine in new bottles. Personalized law brings welfare maximization turbocharged by technology and administered by the state. The endpoint is not radical markets as some modern advocates of the economic analysis of law now propose. The endpoint would be a totalitarian dystopia. Confronting this prospect head on is necessary in order to see the fundamental problems of the personalized law project if executed uncompromisingly and radically. But it is also helpful to identify the limited legitimate uses to which personalized law can be put. I will describe these, and I will also explain why states should focus on regulating “wild” personalization by powerful private actors instead of embarking on an ambitious personalized law program themselves.
I. The “Precision Benefit” Argument
The personalized law project is an ambitious law reform program. It is so ambitious that it can be called revolutionary. The fundamental principle of equality under the law would be substituted by “different laws for different people.” Further, every legal rule in all areas of the law could be personalized. There is nothing in the concept of personalized law which limits its scope. Hence, personalized law is revolutionary both in the sense that its guiding principle would unravel a central pillar of our legal systems, namely equality under the law, and in that it could involve a redesign of all of our laws.
As a revolutionary law reform project, personalized law has a distinctly normative impetus. The program advocates change—radical change. If one advocates such change, one needs to provide a very good justification. “Why personalized law?” is the most important question to ask about the personalized law project. The burden of argument is on the proponents of the project. And it is a tough burden to shoulder given its revolutionary nature.
Ben-Shahar and Porat do not offer a normative justification for personalized law. Rather, they resort to a “technical” argument which seeks to advance the case for personalized law on instrumental grounds. Personalized rules are said to “have the potential to accomplish the underlying goals of any law more effectively” (emphasis added). Personalized law is presented as a precision tool, as a means to better achieve whatever end one wants to achieve with a legal rule (means/ends efficiency). Another way to look at their argument is to say that personalized law would increase legal productivity: doing what you want to do with a legal rule but doing it more productively.
This justification of the personalized law program is surprising. The starting point for the program is that people are different regarding their needs, interests, and preferences and that big data and AI have made it possible to identify these differences much more thoroughly and precisely than ever before in human history. Most would agree with this premise. But then the most “natural” argument to justify personalized law appears to be its potential to massively increase societal welfare by addressing more precisely the individual needs, interests, and preferences of different members of society. To put it differently: personalized law could be a social welfare engine like no other policy instrument before.
So how far does Ben-Shahar’s and Porat’s “precision tool” argument take us? What could give it more force than an argument based on welfarism? A satisfactory answer to these questions cannot be that the “precision tool” argument sounds innocuous and certainly less problematic than an argument based on welfarism. This would amount to sidestepping the crucial issues instead of confronting them. A satisfactory answer could be that the “precision tool” argument is more forceful than an argument based on welfarism if and to the extent the former has a broader scope or, to put it differently, a greater explanatory power than the latter.
Fortunately, the authors’ book is detailed on “personalized law in action,” i.e., it offers a plethora of examples for personalized lawmaking. Reviewing these examples leads me to conclude that there is little (if any) discussion of other goals than efficiency (welfarism or utilitarianism) which could be effectively pursued with personalized rules.
This is most obvious in the authors’ discussion of personalized default rules in contract law and beyond. Personalizing such rules is done to achieve a better satisfaction of individual preferences or needs, i.e., to increase the welfare of those affected by a personalized default rule compared to a non-personalized default rule. The discussion of (mandatory) consumer protection laws yields the same result: “overall welfare increases” with personalization. The authors’ treatment of personalized tort law focuses on “optimal deterrence” as does their analysis of personalized criminal law in seeking to “deter crimes more effectively.” “Optimal deterrence” means minimizing the social costs of civil or criminal delicts—a key goal of welfare economic policy making in these areas. Finally, personalizing voting rights is considered by the authors because “some people value these rights less and refrain from voting” (emphasis added). Hence, individual preferences are best served (satisfied) by a differential allocation of such rights.
True, the authors also discuss whether personalized standards of care in tort law are just in the sense that they are compatible with conceptions of corrective or distributive justice. But that is a claim which is different from the claim that any legislative goal (which?) can be best achieved with personalized law (how?).
Later in the book, the authors claim that personalized protections would serve an “egalitarian goal” better than nonpersonalized protections. But that is a mischaracterization of what personalized protections do. As the authors observe in that section, personalized protections are designed to best meet the individual (idiosyncratic) needs of different individuals. Hence, the goal is to maximize societal welfare by optimally “treating” its individual members. This is the normative individualism of welfare economics. It is (of course) egalitarian in the sense that each individual is treated as a distinct individual. But this is an extremely attenuated conception of distributive justice and at best a side effect but not the goal of such protections.
I am not claiming that there are no legislative goals other than welfarism which could effectively be pursued with personalized law. Surely there are such goals. But one must go beyond the material discussed in the book. An example could be laws to protect the environment and stop or even reverse climate change. Such laws may have other goals than merely to internalize negative externalities (see also infra Part II.B). Personalizing them could help to better achieve these goals. Laws could microtarget climate-sensitive individual consumption patterns such as travel or eating habits with price signals or bans, for example.
What I am claiming is that Ben-Shahar and Porat simply assume that personalized laws could be a “precision tool” to achieve any legislative goal without offering good illustrations to support this assumption. As I have shown, the book’s examples take us to welfarism. More importantly, I am claiming that this is not surprising. To repeat: if the central idea of personalized law is to optimally meet individual needs, interests, and preferences, one almost inevitably ends up with the normative individualism of welfare economics. It is by far the best normative fit for the personalized law project.
Let me offer a “control consideration”: What would be the use of personalized law in a society organised according to Kant’s categorical imperative, i.e., the maxim of generalizability, which states, “act only according to that maxim through which you can at the same time will that it become a universal law”? I submit that the answer to this question is that, in such a society, personalized law would play no role whatsoever. In fact, it would be completely at odds with the spirit of the categorical imperative.
Hence, I conclude that, in essence, the personalized law project, if implemented, will gravitate toward welfare maximization, turbocharged by big data and AI. The “precision benefit” argument is a distraction which does not withstand critical scrutiny. This has consequences.
II. The Specter of Totalitarian Welfarism
A benefit of a welfarist justification for the personalized law project is that it offers a good starting point to explain widespread intuitions about the desirability and legitimacy of personalized law in certain areas and its inappropriateness in others.
Few will dispute that personalizing contract default rules or mandatory protections is a good idea. After all, parties conclude contracts for mutual benefit based on an expected satisfaction of their individual (idiosyncratic) preferences. And few would dispute that it could be sensible to personalize default rules in inheritance law, to give another example. Ben-Shahar and Porat point to the interesting fact that men, on average, bequeath 80% of estates to their widows whereas women bestow only 40% to their widowers. Intestate inheritance law seeks to reflect the putative will (the preferences) of the deceased, and personalized default rules can achieve this goal better than non-personalized defaults.
On the other hand, many will find it inappropriate or even objectionable to personalize constitutional rights or bundles of constitutional rights such as the right to vote. And many would disagree with the authors’ statement that “law too is a type of service” addressing “governance needs” (emphasis added). Can law be reduced to a needs-satisfying factory? Is it not much more than a service?
The reason, I submit, for these intuitions is that welfarism provides only a partial account of the normative principles and goals which are cornerstones of democracies around the world. The normative fabric of our societies is richer and more complex than normative individualism and welfarism make us believe. By focusing on the alleged “precision benefit” of personalized law, Ben-Shahar and Porat lose sight of this complex normative universe. Let me explain in more detail.
A. Welfarism and Rights
Ben-Shahar and Porat’s treatment of constitutional rights appears to be inconsistent. On the one hand, they acknowledge that constitutional rights such as equal protection function as constraints on the personalization enterprise, delineating what it can and what it cannot do. On the other hand, they also subject constitutional rights or bundles of constitutional rights to personalization, as already mentioned. This raises serious questions: Which constitutional rights/constraints are off-limits for the personalization calculation? Which are subject to it, and why or why not?
A convincing account of constitutional rights is to view them as trumps for individuals to veto the unconstrained pursuit of welfarism by majority rule. This is of course the account of constitutional rights famously developed by Ronald Dworkin. On this account, such rights always operate as constraints on the welfarist calculation, and the idea to subject them to this calculation must be rejected. At the same time, viewing rights as constraints on welfarism (and only as constraints) explains the normative intuition about the inappropriateness of personalizing constitutional rights well.
B. Law and Change
Second, the normative individualism of the personalized law program is deficient also because it reflects a static conception of the law and misses its dynamic aspects. Laws do more than servicing current needs and preferences. Laws also express values which reflect what we could be and possibly want to be. Laws seek to influence not only our behavior but also our preferences and values.
Just think of all the new laws worldwide to limit the disastrous effects of climate change such as emission controls, carbon taxes, deforestation bans, etc. Of course, these laws seek to incentivize us to do the right thing (more often), i.e., they send price signals to the markets which have an effect on the supply and demand of desirable or harmful human behavior. But these laws also seek to influence our preferences regarding consumption patterns and, more generally, how to lead a good life.
C. Law and Social Cohesion
Third, humans are social and political animals—zoon politicon as described by Aristotle. We establish our identities as humans through interaction with others. Law is a crucial factor in enabling such interaction. And normative equality under the law as stipulated in the United Nations’ “Universal Declaration of Human Rights” and elsewhere is a key factor for establishing and maintaining social cohesion. It is the glue that binds us together by allowing us to engage with each other at eye level: we enjoy the same constitutional rights and privileges. We are one people.
With personalized law, especially with personalized constitutional rights, this glue disappears. Personalized law, if executed perfectly, amounts to a radical normative individualism. Such radical normative individualism destroys the foundation of our communities and societies. This goes much beyond the coordination problem which the authors discuss. Coordinating traffic when there are personalized speed limits, for example, is a technical problem. Maintaining social cohesion and identity under a regime of different laws for different people is another challenge. Just think of the “constitutional settlement” of extremely controversial issues such as abortion and reproductive rights. The societas would be pulverized if such rights were personalized, I fear.
D. Law and Totalitarianism
Fourth and finally, democratic societies are committed to individual freedom as a central pillar of the constitutional fabric. This extends to the use of personal data. As early as 1983, the German Constitutional Court “invented” the constitutional right of all members of society to “informational self-determination,” striking down a census law at the time. More recently, the European Union enacted the General Data Protection Regulation. A key principle enshrined in Article 5(1)(b) is that personal data shall only be collected “for specified, explicit and legitimate purposes and not further processed in a manner that is incompatible with those purposes . . . (‘purpose limitation’).”
Ben-Shahar and Porat envision a personalized law program which would be run by the state. The guiding principle would not be data minimization and purpose limitation but rather the exact opposite: data maximization and purpose extension. The program would seek to tailor laws to individual needs, interests, and preferences of different members of society—not rough personalization but full or complete personalization. The endpoint would be a government-administered personalization program based on technology-driven data gathering and exploitation. The vision is that of an extremely powerful surveillance and welfare state.
This is an Orwellian prospect. Ben-Shahar and Porat try as best as they can to address critical concerns. But this is a futile enterprise: it is the new paradigm which is highly problematic. The specter of a totalitarian dystopia is not defused by twists and tweaks in the institutional infrastructure. The problems arise from the design: giving states a powerful “precision tool” to maximize societal welfare in a data- and technology-driven planned economy. This is a fundamental threat to individual liberty.
III. Controlling “Wild” Personalization
If the analysis developed above has merit, this implies a significant downsizing of the personalized law project. The project’s normative driving engine is welfarism, or utility maximization. In democratic market economies, welfarism has a role to play. But that role is limited. The pursuit of welfare maximization is constrained by (constitutional) rights, notably the fundamental principles of equality under the law and personal freedom, including in the realm of data production and use. Further, laws also aim to change who we are and what we prefer in a delicate attempt of self-paternalization. In this complex normative space, personalized law will be useful if and to the extent welfarism, constrained by fundamental (constitutional) rights, is a legitimate pursuit. I have attempted to highlight some of these areas in the introductory paragraphs of Part II supra.
The more pressing concern and challenge for states is to control “wild” personalization by powerful private actors. The leaders in the ongoing and accelerating big data and AI revolution are not states around the world but private entrepreneurs, especially in Big Tech. In market economies it could hardly be otherwise. Firms like Amazon, Alphabet (Google), and Meta (Facebook) are creating wealth by making the best possible use of consumers’ data and applying the latest technological tools. Most of this wealth ends up in the hands of the sophisticated technology users rather than with their customers. In fact, the big data and technology revolution has arguably led to consumer exploitation on a grand scale. The focus of state action should not be on designing and running an ambitious personalized law program. It should be on controlling personalization by powerful private actors. The tools to achieve this end are manifold, extending from contract law to antitrust action.
Personalized law is a provocative idea. It challenges a central pillar of our legal systems: equality under the law. In this Essay, I have attempted to demonstrate that the “precision tool” argument advanced by Ben-Shahar and Porat to justify a personalized law program is not convincing. The most natural fit and best possible justification for such a program is welfarism. But a pure welfarist conception of the law is deeply problematic, especially regarding the precarious status of fundamental (constitutional) rights in such a conception. I have sought to demonstrate the limited role of personalized law in a more complex and balanced normative conception for law reform, and I have argued that the focus of state action should not be on running a personalized law scheme but on controlling wild personalization by powerful private actors.
This might be considered sobering news for the personalized law project. But, as already explained in the Introduction, the project’s promotors themselves are not fully committed to realizing it—at least not full-scale. The “brave new world” they unfold before our eyes would not be a world for humans but for machines—and for machines only. It is almost comic to read the story of a day in the life of David and Abigail which is shaped by sophisticated technological applications and personalization gimmicks. Should we really believe that it is humans as we know them who are populating such a world, driving cars themselves, going to physical stores, receiving physical letters, etc.? It reads as if super-advanced technology has been implanted into a museum, where everything else has been conserved, i.e., humans as we know them, traditional ways of living, etc.
I submit that if and to the extent personalized law becomes a noticeable reality, humans will increasingly depart from this reality. The authors themselves acknowledge that humans might need machines to help them understand personalized commands and comply with them. But surely this would only be the first step. It would be much more efficient to eliminate humans as real actors altogether. In a fully personalized world, most of laws’ commands would not be directed to humans but to machines.
This is a fundamental challenge for the personalized law project because the whole idea of personalization, i.e., to take humans as individuals seriously, becomes obsolete in a world of smart machines. Machines would make sophisticated rules for . . . machines. Everything would run smoothly without human input or involvement. Rather than leading to a more humane world by taking each and everybody more seriously, personalized law would eventually dehumanize our societies. The revolution would consume its own children. One can start speculating what the role of humans could be in such a world. Would they populate a delightful virtual reality created by machines?
* * *
Horst Eidenmüller is a Statutory Professor for Commercial Law at the University of Oxford and a Professorial Fellow of St. Hugh’s College, Oxford. He is also a Member of the Berlin-Brandenburg Academy of Sciences and Humanities (BBAW) and a Research Member of the European Corporate Governance Institute (ECGI).
* * *