Personalized Law
Different Rules for Different People
OMRI
BEN- SHAHAR
Leo Herzel Professor of Law University of Chicago Law School
ARIEL PORAT
Alain Poher Professor of Law and President
Tel-Aviv University
3
Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and certain other countries.
Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America.
© Oxford University Press 2021
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by license, or under terms agreed with the appropriate reproduction rights organization. Inquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above.
You must not circulate this work in any other form and you must impose this same condition on any acquirer.
Library of Congress Cataloging-in-Publication Data
Names: Ben-Shahar, Omri, author. | Porat, Ariel, author.
Title: Personalized law : different rules for different people / Omri Ben-Shahar, Leo Herzel Professor of Law, University of Chicago Law School and Ariel Porat, Alain Poher Professor of Law and President Tel-Aviv University.
Description: New York, NY : Oxford University Press, 2021. | Includes bibliographical references and index.
Identifiers: LCCN 2020055609 (print) | LCCN 2020055610 (ebook) | ISBN 9780197522813 (hardback) | ISBN 9780197522820 (updf) | ISBN 9780197522837 (epub) | ISBN 9780197522844 (online)
Subjects: LCSH: Law—Philosophy. | Obligations (Law) | Liability (Law) | Human rights. | Equality before the law. | Individual differences—Government policy.
Classification: LCC K 258.B46 2021 (print) | LCC K 258 (ebook) | DDC 340/.1—dc23
LC record available at https://lccn.loc.gov/2020055609
LC ebook record available at https://lccn.loc.gov/2020055610
DOI: 10.1093/oso/9780197522813.001.0001
1 3 5 7 9 8 6 4 2
Printed by Sheridan Books, Inc., United States of America
Note to Readers
This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is based upon sources believed to be accurate and reliable and is intended to be current as of the time it was written. It is sold with the understanding that the publisher is not engaged in rendering legal, accounting, or other professional services. If legal advice or other expert assistance is required, the services of a competent professional person should be sought. Also, to confirm that the information has not been affected or changed by recent developments, traditional legal research techniques should be used, including checking primary sources where appropriate.
(Based on the Declaration of Principles jointly adopted by a Committee of the American Bar Association and a Committee of Publishers and Associations.)
You may order this or any other Oxford University Press publication by visiting the Oxford University Press website at www.oup.com.
To my mother, Yael, and her lifelong defiance of uniformity
Omri Ben-Shahar
To Timna, my love
Ariel Porat
Preface
This is a book of science fiction. Or maybe not.
Personalized law is a vision of a brave new world, the law of the future, where people will each be subject to different legal rules, each bound by their own personally tailored law. The data and technology to run personalized law at full throttle are not currently available, but embryos of this regime are sprouting in the background. If technically feasible, do we want it? Does personalized law harbor a utopic promise similar to the diagnostic and treatment powers of personalized medicine? Or would it produce alienation, demoralization, and discrimination?
How did the two of us, card-carrying scholars of private law and lawand-economics, end up writing this book, on the jurisprudence of personalized law?
In 2014, we each separately published ideas in deep disagreement. One of us (Ben-Shahar) was a co-author of a book arguing that mandated disclosures and default rules are regulatory techniques that consistently fail and cannot be fixed.1 The other (Porat) was the co-author of a law review article arguing that mandated disclosures and default rules would work well if personalized.2
For a couple of years, we feuded—could personalized tailoring of defaults and disclosures really make such a difference—until we decided to examine personalized law in other contexts. We chose two settings, representing two of the more powerful forms of legal intervention: tort law’s standard of care and statutory consumer protections in contract law. We wrote two articles, one on each of these legal techniques. There, we began an enterprise that culminated in this book, comparing the social impact of personalized rules visà-vis our existing uniform rules. We asked: What if each potential injurer were faced with a different standard of care? What if each consumer received a different bundle of mandatory protections?
1 Omri Ben-Shahar & Carl E. Schneider, More Than You Wanted To Know: The Failure of Mandated Disclosure (2014).
2 Ariel Porat & Lior J. Strahilevitz, Personalizing Default Rules and Disclosure with Big Data, 112 Mich. L. Rev. 1417 (2014).
We were struck by the richness of insights these comparisons yielded. We started to see a common theme, one that we suspected would run across every area of law. Personalized law could distribute commands with more precision, which means that the goals of the law—any goals of any law—would be more fully promoted. We also began to see the challenges and problems that could undermine these advances. Personalized law, we realized, is a fertile platform to offer new angles on basic questions of law, equality, coordination, and an array of social policies. It raises challenging questions on the relationship between law and Big Data.
We marinated for a couple years in the plans for this book. We presented our ideas to various audiences, listened to various objections, relished the interest the topic evoked among people, and continued to examine effects of personalized law in other legal areas. We were ready to begin writing a manuscript, when an unexpected challenge befell us: Porat was elected President of Tel-Aviv University. Remaining committed to this book, our intellectual partnership now had to endure late night overseas conversations and stretches of delay.
Numerous people have contributed to our thinking about personalized law. Primary among them is our colleague and friend Lior Strahilevitz, who co-wrote with Porat the first article on the topic. We also acknowledge in the footnotes the writings of researchers who examined aspects of tailored legal institutions, and we ought to mention the creative work of our colleagues Anthony Casey and Anthony Niblett, who separately developed ideas about data-driven micro commands. In addition, we have presented versions of our work to audiences at our home institutions, the University of Chicago and Tel-Aviv University, and to conference and workshop participants at the German Advisory Council for Consumer Affairs, the Society of European Contract Law, Stanford University, University of Illinois, the Symposium on Granular Legal Norms at Villa Vigoni, and Yale Law School. Many participants in these events contributed critical comments. We owe special thanks to Horst Eidenmüller, Daniel Hemel, Don Herzog, and Carl Schneider, who read the entire manuscript and generously shared with us critical, and at times biting, commentary. We are also grateful to Will Baude, Ziv Ben-Shahar, Sarah Clarke, Lee Fennell, Aziz Huq, Robin Kar, Emma Kaufman, Jared Mayer, and Haggai Porat. Finally, we benefited from superb research assistance by Kathryn Banks, as well as additional very able research help from Brenna Darling, Brigid Larkin, Sahar Segal, Garrett Solberg, and Tom Zur. Ben-Shahar gratefully acknowledges the financial support of the Daniel and Gloria Kearney Fund and the Coase-Sandor Institute for Law and Economics.
1 Introduction
We live in a world of one-size-fits-all law. People are different, but the laws that govern them are uniform. From traffic commands to license requirements, from legal protections to legal procedures, people’s rights and duties are presently written and intended to apply identically to all. Everyone must satisfy the same criteria and cross the same thresholds to qualify for any legal treatment.
Personalized law—rules that vary person by person—will change that. A legal norm calibrated for the “reasonable person” can be replaced by a multitude of personalized rules, each person with their own “reasonable you” standard. Better drivers would be free to drive faster, more skilled doctors would be held to higher standards of care, vulnerable employees would receive stronger protections, age restrictions for the young and the elderly would vary according to individual competence, and each borrower would be entitled to a personalized loan disclosure tailored to their specific concerns and delivered in a format fitting their mental capacity.
Uniformity of the law is so universally accepted that it is little remarked. When the age of twenty-one is the legal minimum for purchasing alcohol, it is twenty-one for all. If lawmakers conclude that mortgage borrowers need additional protections—for example, a right to withdraw, a limit on a certain hidden fee, or a warning—they bestow the protection in equal measure on all. Personalized law would set different age thresholds across people, allowing some to engage in the activity earlier than others. It would give different bundles of protections to consumers, based on their specific needs. As information becomes available about relevant differences between individuals, like who poses higher risk, who might be harmed more, or what personal goals they have, the legal treatment could accommodate such diversity.
Personalized law is the name we give to a novel institutional template that merges a jurisprudential transformation of to each their own law with an ambitious implementation project relying on Big Data. Tailoring individual rules requires information about people and a method to determine which personal attributes are relevant to the design of their personalized legal
treatments. The raw information is increasingly available, and algorithms to determine which bits of it are relevant to the calibration of optimal legal commands are increasingly reliable.
The digital revolution created databases with rich personal information used to personalize other sectors and other treatments that were once uniform. Could law be next? Like personalized medicine or education, personalized law could bring enormous social benefits. But unlike other sectors that use data and predictions to customize the treatments of different people, personalized law raises tough questions about equality and the rule of law, strategic manipulation and social coordination, the sources and quality of data, and the substitution of human discretion with artificial intelligence (AI).
Personalized law is a timely, almost inevitable inquiry. A fantasy only a short while ago, the technology to roll it out in some legal domains may be at our doorstep. Should we embrace it as an exciting breakthrough that will finally allow us to treat humans as individuals, not as a population, or reject it as a chilling, Big Brother dystopia? Is it possible, as this book suggests, to design a legal system that delivers the benefits of personalized law with the right constraints and safeguards?
Our interest is not in the law of Big Data or AI. The rise of digital technologies has created previously unthinkable products, organized new markets and communities, and affected our daily lives in profound ways. Data have grown to be the fuel that runs the new economy. It is a revolution that has taken center stage in law, posing novel challenges to assure that the digital ecology, with all its good, does not cause unintended harm. Who owns the information, how may it be used, and what limits on data collection to install are the central questions of the law of digital data.
Personalized Law reverses the inquiry. Rather than ask how the law should govern and safeguard the Big Data enterprise, our book poses the opposite question: how data and AI might govern and sharpen the tailoring of legal rules. Could the resulting algorithms augment and replace human discretion in identifying “optimal” legal commands? Could they be instructed to find relevant differences across people and use these differences to personalize sanctions, rights, and duties?
This book is not the first to suggest that robotic methods could improve law, substituting for humans in guiding behavior and resolving disputes.1
1 In a parallel inquiry, two articles explored the use of algorithmic methods to the design of legal commands. See generally John O. McGinnis & Steven Wasick, Law’s Algorithm, 66 Fla. L. Rev. 991 (2014); Anthony J. Casey & Anthony Niblett, The Death of Rules and Standards, 92 Ind. L.J. 1401
Novel computational schemes germinating in various areas of the law are entrusted with tasks traditionally performed by judges and lawyers, and carry them out with surprisingly increasing competence. Shadowed by colossal concerns, early experimentations with innovations like algorithmic judges are sometimes encouraging. Courts, for example, are using prediction tools to assist with criminal sentencing and parole decisions, managing to eliminate some effects of judges’ errors and biases, but potentially introducing new distortions. Less controversially, legal advice is now commercially available from algorithms. Plug into a digital interface some basic information about your dispute, and within nanoseconds you will receive a fairly accurate prediction, based on machine analysis of past cases, of how you will fare in court. The growing use of such tools renders settlements effortless, litigation unnecessary, and attorneys obsolete. And yet a major challenge for this embryonic enterprise is the quality of the data. When prediction algorithms are trained with information reflecting past discrimination and biases, they may reproduce and perpetuate them.
Hopeful as many of us may be about data-driven dispute resolution tools, rule-making is a different ballgame. Dispute resolution is a problem of examining evidence and applying the instructions embedded in existing rules. Lawmaking is a deeply normative project of issuing rules to guide people toward desired behavior. The rules are enacted to advance social goals and resolve difficult political and ideological trade-offs. How could these be left to machines? Moreover, people differ in so many ways; who is to decide which differences matter when tailoring different rules to different people? These questions must be answered if the personalization of rules is to depend on data technology, and we dedicate much of this book to them.
Personalized law is a revolutionary jurisprudence, but it is also remarkably not new. The idea that legal commands should be customized to the idiosyncratic characteristics of individual actors has been with us for millennia. The Fifth Commandment—“Honour thy father and thy mother”—requires significant amount of nuanced personal context in application. It calls for obedience, material support, and personal warmth, but the specific requirements vary between people. “A person’s wish is the way to honor him,”
(2017). Casey and Niblett in particular explore the use of “microdirectives” derived by algorithms and tailored to every possible scenario. However, their work does not examine personalized law and focuses instead on how machine-derived rules might be created and communicated.
says a medieval source,2 and a voluminous interpretive commentary has labored to demonstrate which factors matter when, and to provide guidance for personalized application.
Presently, tort law, for example, instructs courts to assess to each victim a compensation equal to their idiosyncratic loss, and criminal law tailors each offender’s sanction based on a personalized recidivism risk assessment. These flickers of personalization are notable exceptions to a background of uniformity. The type of personalized law this book introduces is broader, both in the factors it relies upon and the contexts to which it applies. To imagine the potential breadth of personalized law—the brave new world this book invites you to consider—join us for a brief tour through “a day in the life of David and Abigail.”
Diary: A Day in the Life of David and Abigail
David and Abigail begin their day with a morning exercise and stretch routine. They each get into their smart training outfits, which transmit their fitness activity to their phones. As their phones broadcast real-time recommended workouts, they can also post workout summaries online. David’s scores are posted automatically on Facebook. Privacy laws, which used to require prior consent to such sharing, now allow people to opt into a “predicted consent” regime that anticipates their preferences. Based on his past sharing practices, David is predicted to agree. Abigail’s workout summary is not shared on social media, as she never opted into “predicted consent” and is governed by a stricter privacy default rule.
A strong cup of coffee after exercise triggers David’s severe heartburn. His physician prescribed a new drug for this condition, and David decides to hop over to the pharmacy to purchase the cure. As he drives out into the street, the dashboard indicates a required maximum speed of 15 mph. A day earlier, when turning into the same road at the same time of the day in similar traffic conditions, his speed limit was 25 mph. David smirks; it was undeniably a (Fitbitrecorded) night of turbulent sleep.
At the pharmacy counter David receives his pills. There is no paperwork; but his smartphone instantaneously flashes a warning, mandated by law: “You must wait one hour after consuming this drug before taking you daily dosage
2 Sefer Hasidim § 152 (Ger.) (c. 1101–1300 c e ).
of insulin.” A second warning then adds, “Do not take the drug while you are experiencing a migraine.” Indeed, David takes insulin every day, and unfortunately suffers from occasional migraines. He remembers anecdotally that when Abigail brought home this drug the year before, she received different warnings. In fact, she was recently warned to stop using the drug on account of her pregnancy—a warning she received before she had found out she was pregnant, instantaneously after she purchased a pregnancy test
From the pharmacy, David drives to Office Depot to purchase a new printer for his home network system. Now, the speed limit display flashes 20 mph—is it the cup of coffee he bought at Starbucks while waiting for the prescription? At the store checkout, a $25 surcharge, reflecting two separate premiums, is added to the price of the printer. The first, a $5 add-on, reflects the longer-than-average right to withdraw automatically granted to David by law, which he greatly appreciates, as he typically shops without knowing much about electronics and needs time to verify that the purchase is worthy. The second, a $20 premium, is the cost of an extended implied warranty covering all problems with the device for two years. The law now provides different echelons of implied warranties, that each vendor is permitted to separately price, and when buying technological products David is automatically fitted with the most generous warranty available.
Upon returning to his car with the new product, David finds a parking ticket on the windshield of his Acura. The fine is $97. An acquaintance mentioned to him a few days earlier that he received a parking ticket of $39 in the same neighborhood. Hmm, David didn’t realize that his friend’s income was so much lower. He makes a mental note for next time to check in advance the personalized fine for unpaid parking. His ParkChicago app has a button that flashes this information. Better yet, he instantly changes the app’s settings so that he will receive this fine information notification automatically, every time the app recognizes that the car is parked in the street.
David returns home, and is now browsing the web looking to reserve a rental car for an upcoming skiing trip. He is informed that Colorado requires him to purchase auto liability insurance of at least $75,000. He restarts the reservation, this time under the name of his skiing partner, and now to his delight the statutory minimum is only $40,000. The partner, a friend from work, is older and earns quite a bit more than David. Later, David books a flight on Southwest.com for a family vacation. He buys three tickets. The toddler’s fare is non-refundable, Abigail’s fare allows for cancellation within seventytwo hours with a full refund, and his own fare allows for a full seven-day
cancellation period. Each ticket is priced differently. The law nowadays mandates the refundability rules, but the mandated minimum duration varies across passengers. “They have to fix the algorithm,” David muses. “I do need flexibility in my business travel, but what is the sense of differential rights when we travel as a family?”
While on his computer, David checks his social network page. He is no longer receiving news reflecting his personal interests, he contently notices. As happens every year, he was recently “asked” in a data preferences survey administered by the Federal Trade Commission to rank various choices on a scale of 1 to 5. One of the questions related to preferences for personalized news notifications. Others applied to digital use of his search and content history for various sharing and advertising practices. He chose Level 5 (no use of personalized data) for news delivery and Level 1 (use any data other than information about minors) for most of everything else.
Mail arrives. A notice from Geico tells David that the auto policy is up for renewal and that a 14% reduction in the premium is calculated to reflect a lower expected liability. He logs in to the Illinois Department of Motor Vehicles website and sees that his personal driving “risk-score”—a measure reflecting each driver’s accident riskiness—is lower than it was last week, finally falling within the “green” range of the diagram. This risk score is one of the factors the state uses to determine the conditions for license renewal (“green” means automatic renewal). Insurers piggyback on the score to predict a driver’s tort liabilities for future accidents and price auto liability insurance accordingly.
A letter from his credit card company informs David that the fraudulent charges on his lost Visa card were removed. It also tells him that he will be responsible for $200 of the charges—the maximum personalized liability that federal law permits in his case. When he lost his card last year, he was liable for only $100. “Bummer,” he thinks, remembering the days of the one-size $50 cap for liability. “The new personalized limits caught up with my sloppiness in handling my cards.”
Dinner at a restaurant. David and Abigail each view the menu on their linked devices. They see the same menu items and the same prices, but each sees different nutritional data (mandated by law). David’s blood sugar responds poorly to carbohydrates, and thus the carb score is labeled prominently next to each menu item on his display. Abigail, who is pregnant, is avoiding added sugars, so the sugar content information is uniquely labeled for her. (Before her pregnancy, she was on a low-fat diet, and at the time that information automatically appeared on her personalized menu display.)
At dinner, the couple discuss the pregnancy and their finances. They ponder how the birth of their second child will affect their estate plans and the allocation of the estate if one of them dies. They never wrote wills, but Abigail checks the Wills.gov site and enters the couple’s social security numbers. A webpage opens and shows that the family growth will not change the allocation of David’s estate: the entire estate will still go to Abigail, and nothing will go to the kids. For Abigail, the website shows that the birth will change the allocation, and her estate would be divided equally between David and the two kids, with one-third going to each. “No need to write a will,” they grin
Dinner ends and the valet brings over their car. David takes the driver’s seat. The ignition will not start. “Aghhh,” he moans, “they know I drank alcohol tonight.” He blows into the breathalyzer now mandated to be installed in every car. His blood alcohol count reads 0.08%, above his personalized legal limit of 0.05%. Abigail, who drank only a sip of wine, takes the driver’s seat, logs in biometrically, breathes her 0.02% alcohol count into the device (her own personalized limit is 0.03%—the law doesn’t adjust it during pregnancy), and gets the green light from the obstinate ignition control. “Let’s go home,” she says.
The diary of David and Abigail may send chills up your spine. This, even before we get to the really big things. David and Abigail’s day was full of largely inconsequential legal matters. They did not take out a mortgage, sue or get sued, commit or suffer a crime, or buy a car. They did not change their employment, get married, or suffer a tort injury. They did not purchase insurance, open an investment account, or contract for home repair. The law stayed in the background. Their environment, however, was shaped by an extensive personalized umbrella of legal rights and duties, each with their different commands.
Personalized law can, we hope to show, deliver social benefits: fewer accidents, better protections, less incarceration, more efficient trade, improved medical decisions, easier dispute resolution, and lower transaction costs. But do you want to live in this world? Where so much that’s known about you is used to formulate your own personal legal space? Where other people are allowed to purchase alcohol at a younger age than you, drive faster, or receive strong consumer protections? There is much comfort in collective treatment. But there is also dignity in being treated as an individual, not a population.
The basic question this book wrestles with is the right scope of personalized law. Is it a regime suited for the occasional modest application, striving to incrementally improve unambitious legal rules, like mandated disclosures?
Or is it a radical transformation that could gradually sweep much of the law? We ask, and identify factors that might help us answer, whether the glue that holds social coordination together—equality before the law—can survive broad applications of personalized law.
One of this book’s methods is to explore the scope of personalized law along two dimensions—the extensive and the intensive. We already gave examples of potential applications: traffic regulations calibrated to each driver’s riskiness, personalized warnings and disclosures, parking and criminal fines that depend on income, age-of-capacity rules reflecting actual individual capacities, and default rules that vary between people. We plan to provide many more applications throughout the book. Yet we are acutely aware that the extensive dimension—which rules and legal areas could or ought to be subject to personalization—will not be fully resolved here. We hope, at the very least, to identify legal areas that would be good candidates to personalize first, and some plausible followers.
A second aspect of personalized law to which we pay close attention is what we call the intensive dimension—how many factors are used as inputs for each legal rule. When we say “factors” we refer to individual characteristics that would be used to differentiate people when prescribing personalized commands. Personalized law can be done intuitively, based on a handful of factors, as when judges focus on a few salient characteristics to assess a criminal defendant’s risk of recidivism, or when they elevate the standard of care for a skilled physician charged with malpractice. But personalization can be done far more effectively, if based on Big Data and algorithmic analysis, in the manner that, for example, auto insurance companies predict policyholders’ accident risks. For such intensive application, we will have to answer a plethora of questions. Where will these data come from? How will the different factors be given appropriate weights? What if the data are incomplete? or biased? or out-of-date?
As we wrestle throughout the book with the extensive and intensive scope challenges—where and how to apply personalized legal commands—we hope to provoke more than just curiosity in our readers. Our goal is to examine and shed the most favorable light on personalized law, without perpetrating advocacy. We want to take the basic idea—that individual traits matter for the tailoring and application of legal rules—which is obviously right, and explore the challenges it faces when stretched to its limit case.
To that end we have to accomplish several tasks. First, we have to show the benefits of personalized law. We dedicate the first part of this book to
that task. Second, we must give at least some guidance on how personalized law would work in action. We do that in Part II of the book by examining personalized rules in specific legal areas. And third, and perhaps most critically, we have to anticipate the problems of personalized law and the social values it threatens. Indeed, throughout the book we present an expanded and sometimes radical vision of personalized law, both on the intensive and expensive fronts, in order to bring to light these problems. The second half of the book—Parts III and IV—identifies the primary objections and evaluates them. Let us briefly mention the most alarming aspects of personalized law that we later discuss.
Personalized law is alarming because it gives the government immense data about citizens. Disturbed as some may be about the ongoing commercial enterprise of collecting and using personal information for profit, this anxiety deepens when the databases are in the hands of governments unchained from the constraints of treating all citizens alike.
Personalized law is alarming because it threatens a fundamental principle of liberalism: equality under the law. True, people are different, but are their differences relevant enough to justify individualized legal treatment? Personalized law would need to use factors like age, income, experience, and physical and cognitive abilities to tailor individualized rules, and it would be all the more alarming if it also uses, or has a differential impact along, sensitive classifications like race, religion, sex, or gender identity. Are such distinctions even permitted under our constitutional regime? Could they be avoided?
Personalized law is alarming because it might destroy social coordination. Would traffic flow safely when each driver is subject to their own personal speed limits and rules of the road? Could litigation be managed fairly if litigants vary in their procedural rights? Markets and trade have delivered enormous benefits because of the standardization of products, prices, and legal rights. Personalized law would undo the uniformity of these tenets, threatening the flow of commerce. Moreover, uniformity breeds conformity: people acquiesce to legal commands that apply to all. Would personalized law breed resentment and disobedience?
Personalized law is alarming because legal treatments depend on individual characteristics, many of which could be manipulated. If people receive better consumer protections when they are thought to have greater needs, they might invest less in reducing these needs. If people who improve their human capital are struck with more demanding standards of care, this may chill their incentive to improve. With so many opportunities to up- or
downgrade one’s rights and duties, wasteful maneuvers to qualify for more favorable treatment could become a major distraction.
Personalized law is alarming because it delegates rule-making to nonhuman devices. Competent as the algorithms might be in executing well-programmed instructions, should law be reduced to lines of code? Algorithms could be told to “optimize” the commands, but they have to be served with a list of objectives. How do we quantify and weigh the competing social objectives of any given law? Tort law seeks to promote deterrence and corrective justice; criminal sanctions have the additional goals of retribution; and consumer protections try to protect people and improve markets. When laws are enacted to advance multiple and sometimes conflicting goals, algorithmic execution of personalized law would need a program that distills these into a mathematical objective function, relying on lawmakers not only to list the goals, but also to determine in advance the exact relative importance of each. Lawmakers have never before been asked to do this.
Personalized law is alarming for such grave reasons that it might seem to be a very bad idea, even a non-starter. It is no secret that the alternative—onesize-fits-all rules—has its own shortcomings. This is why legal commands are already customized, tailored to apply differently in different circumstances and, in subtle ways, even to personal circumstances. But the existing templates of legal granularity fall far short of personalized law. It is tempting to assume that a sweeping move to personalized law would be so costly, so impossible to do well and so contrary to our shared notions of legal order, that it would be better to maintain uniform laws and fix their distortions and imprecision through less dramatic interventions.
We devote a big part of this book to discussing these alarms, but we start it with an account of the benefits of personalized law. In fact, these benefits are so obvious and plausible that we don’t need to spend too much space to demonstrate their potential. Plain and simple: compared to uniform law, personalized law promotes more effectively the law’s underlying goals. Any goals, of any law. If tort law seeks to deter unsafe behavior, personalized commands would reduce both accidents and prevention costs. If its goal is to compensate victims according a specific make-whole principle, personalized rules would guarantee more accuracy in the assessment of damages. If the goal of default rules is to mimic people’s will and reduce the cost of expressing their desires, personalized defaults could do better than majoritarian ones by recognizing and giving effect to the heterogeneity of preferences. Personalized law will award licenses to those who truly merit them; allocate property rights or
permitted uses to more deserving users; and provide guidance to people who engage in dangerous activities on the basis of their personal, rather than average, fit. Personalized law might even accomplish a miracle and make what is arguably the least successful of all regulatory techniques, mandated disclosure, work—by giving people only the information useful to them, at the time that fits their schedules and in the format that best suits their cognitive capabilities. Across all of law, personalized rules could achieve better results with less adjudicative error, ultimately improving access to courts and justice.
Personalized law does all that because it differentiates legal commands on the basis of relevant circumstances and thus avoids the inaccuracies and unfairness of crude uniform treatments. This feature—higher precision is the same reason that any personalized scheme is potentially superior to its one-size alternative. It is why custom-made shoes fit better than single size, why personalized nutrition plans produce more effective diets than standard meal plans, and why medicine based on personalized diagnostics cures better than one-size-fits-all treatments. Uniformity, even if optimal on average, creates a poor fit for a lot of non-average people in a population with diverse preferences, characteristics, histories, and means.
There are alluring aspects to a world of personalized law. Some of them could be thought of as “technocratic”—convenience, efficiency, and precision in the administration of legal rules, the avoidance of errors, and the achievement of better social order. Information and statistical correlations replace guesses and intuition. Algorithms trained to detect patterns in the data replace humans who rely on impression and faulty cognition or memory. These are engineering achievements, but what about humanism, sympathy, and intuition?
Do not be misled by the mechanical aspects. Personalized law, we dare suggest, has the potential to advance, in a novel way, the administration of justice and lenity. At its core, personalized law is the utmost manifestation of the centrality of an individual, not the colony. In our view, it is the alternative regime—uniform law’s basic principle of packing heterogeneous people into the same template, no matter how different these people are—that elevates administrative technocratic values over fairness. In so many areas outside the law, the idea of treating people like individuals and recognizing their uniqueness is the foundation of dignity and autonomy. Personalized law has the potential to reinforce this conviction.
The genius of personalized law is its marriage of technical-scientific method with broad-minded goals and aspirations. Consider consumer
protection law—one of the hallmarks of progressive regulation. Laws in this area typically begin with an account of why people need protection, identifying the market risks they face when unprotected. Often, such accounts point to the cognitive limitations of consumers, and they tailor solutions to address those limitations. How crude! Cognition is the most personal of characteristics, one that varies so richly across people, and yet our laws currently treat people as if they are uniformly cognitively limited. Consequently, by handing out the protections one-size to all, these laws dilute the treatment to those most deserving.
Instead, to understand the different challenges people face and protections they need, we could use data at every step. Personalized law would then distribute the protections to advance the normative aspirations of the protective regime. Rather than give all consumers the same rights, let us identify those whose need for the protection is particularly acute and award them the greatest quantum of rights. Personalized law would also identify people whose needs are low, or who are predicted to abuse the system, and award them the least measure of protection. In between, people will receive more or less protection, along a continuum, in accordance with their attributes thought to be relevant to the goals programmed into the algorithm. In short, the technical aspects of personalized law are put in place to guarantee that the non-technical goals are optimally served.
Finally, a word about what personalized law is not, and then a word about what this book is not. Personalized law is not a social “credit” system to evaluate people’s trustworthiness and loyalty. People will not be barred from airports, rejected from schools, or refused occupational licenses on the basis of political views and civic morals. On the contrary, personalized law is a scheme intended to increase participation and reduce unnecessary barriers. Some people in some areas might have to meet more exacting duties because they pose higher specific risks. Simultaneously, countless other people will enjoy less burdensome duties and greater access. Personalized systems can become abusive when the goals driving them are wrong. They can be liberating when the goals are worthy.
This book is not a manual for reform. It examines an extreme scope of personalized law, and readers might be troubled by this radical vision and by the assumptions that underlie it. Some would be agitated by the massive information personalized law wants governments to have, and the trust it wants to put in government-run algorithms. Others would resent various substantive applications, or how personalized law might classify people. It is important
to state in advance that the model presented in this book is a methodological benchmark. Throughout the book we identify the dimensions along which personalized law could be dimmed down. We will argue that some degree of personalization of legal commands is justified when it can achieve its purposes. In the end, the incidence and magnitude of personalized law must be determined by convincingly showing that the specific application is beneficial, implementable, and immune to the overhanging problems.
The Plan of the Book
Personalized law is a novel institution, so a good place to begin is by explaining what it is. Chapter 2 describes a model of personalized law in relation to several comparative baselines. It is a form of customized law, but it is one that focuses on factors internal to each actor, rather than on their external circumstances. It is a new template, but with close resemblance to long-standing traditions of personalized rules both in law and in extralegal systems (like religion and private regulation). It is based on Big Data and algorithmic analysis, which pose entirely novel challenges for the legal process.
One cannot understand what personalized law is without a full sense of what it aspires to accomplish. Chapter 3 lays out the basic argument— that personalized law could better fulfill the goals of any system of rules. Personalization is common in many areas—so transformative of sectors that once treated people uniformly—that we must ask why not in law? Chapter 3 looks at other sectors and examines how the adoption of personalized modules improved their performance. It exposes not only the advantages of personalization, but also how inevitable it has become in our data-fueled society. When the template is adapted to law, we see that, regardless of what the objectives underpinning any specific law are, they can be advanced more adequately if people are treated as if they are not interchangeable.
The rest of the book then uses a variety of delicate brushes to depict the manifold issues that the abstract precision claim in Chapter 3 ignores. Part II of the book continues to build the case for personalized law by demonstrating how it would apply to a variety of legal areas. We did not want to write a painstaking survey that marches our readers through all of law, counting the endless applications of personalized rules. Instead, we present three prisms, each slicing the legal system along different cross-sections. The first, in Chapter 4, is a set of doctrinal case studies: the common law duty to
