Adam Davidson1Adam Davidson is a Harry A. Bigelow Teaching Fellow and Lecturer in Law at the University of Chicago Law School. He thanks Omri Ben-Shahar, Ariel Porat, and the participants in the Personalized Law Symposium for their discussion and suggestions. He also thanks Aneil Kovvali and Elizabeth Reese for their suggestions on an earlier draft and the University of Chicago Law Review Online editors for their work on the piece.
* * *
A part of the series, Personalized Law.
I. Personalized Law and the Dangerous Few
Professors Omri Ben-Shahar and Ariel Porat paint a fascinating picture of a potentially very different legal future in Personalized Law: Different Rules for Different People. In many ways, this future seems utopian. Finally, the law will not have to make rough cuts in society but can instead treat each person as the individual they are. But to do this, the state must access—whether voluntarily or through coercion—a dystopian amount of information about each person. For personalized law to truly live up to its promise, few if any people could opt out of this data collection, as personalized law requires both large amounts of data from large numbers of individuals in order to operate effective algorithms and large amounts of data from each individual in order to make accurate predictions and recommendations about how the law should apply to that individual. At the moment, this seems like a minimal hurdle. Few people object, and even fewer truly opt out of, the trove of data that is collected from us. When, for example, was the last time you decided not to visit a website because you did not want to accept the tracking cookies it warned you about? But the shift from private to state, along with the expansion of the types and amount of data collected, may cause significant pushback.
I note that it is not clear how, or whether, public reaction would differ if the state collected data directly (as it does when we file taxes, for example), or if it gained access to data through what would essentially amount to public-private partnerships in which private entities collected our data and then gave it to the government. Borrowing from Ben-Shahar and Porat, I assume for this Essay that whichever path the government took, it would collect this data openly instead of surreptitiously because of the importance of transparency in reducing the negative externalities from a personalized law regime. However the government gets the data, the end state of the world is the same: the government would have and use our data to create the personalized law regime.
Despite these objections and qualifications, there is one area where there is seemingly universal agreement that increased data collection and lessened privacy interests, even if mandatory and state-sanctioned,2I note here that I am using the term “state-sanctioned” loosely. Transformative justice practices often explicitly reject the involvement of the modern state. This Essay assumes, however, the possibility of a future in which there is an abolitionist “state.” I will therefore refer to the “state” as though it is synonymous with the sorts of community-oriented solutions that transformative justice praxis calls for. are justified. That is when attempting to address people who have committed criminal, or otherwise sufficiently great, harm. In this space, everyone from the most tough-on-crime prosecutors to ardent carceral abolitionists recognizes that in order to ameliorate the harm a person has done, that person may need to relinquish certain privacy rights. In the carceral space, this is most obviously seen through the greatly diminished privacy rights of imprisoned people. But restorative and transformative justice practices too often require participants to share information about themselves that they would otherwise prefer to keep secret. Indeed, this full accounting seems necessary to both the personal and community repair that restorative and transformative justice practices seek to bring about, and it is why there has been a push to recognize information shared within restorative justice spaces as privileged.
This agreement around, and the necessity for, involuntary data collection and use is further heightened when dealing with the category of people whom abolitionists call “the dangerous few.” Who these dangerous few are, how many they might be, and how they are to be identified, is a matter of much debate. But the basic idea of the dangerous few is not complicated. There may be, most abolitionists concede, some number of people who are so dangerous to society—so likely to continue committing serious harm even in an abolitionist world—that they must be incapacitated in some way. These “dangerous few” are people who, despite the advances of an abolitionist world, would continue to do harm to others unless they are restrained in a way that is carceral in nature. As Professor Allegra M. McLeod explained, the dangerous few “are only those who are intent on perpetrating acts of vicious harm against others such that they are an imminent threat to all those around them regardless of their circumstances.” While Professor McLeod lists numerous, indeed millions, of currently imprisoned people who are not part of the dangerous few, she, like other abolitionists, does not identify who the dangerous few actually are. This is unsurprising. Many people once thought incorrigible, indeed even those who have been sentenced to death, have turned into positive members of society. But where humanity’s intuition-based and often biased predictions of who the dangerous few are have failed, personalized law’s focus on algorithmic prediction could succeed.
This Essay proceeds as follows. Part II will describe how personalized law, through a combination of the inherent technological advances necessary for a personalized law regime, widespread cultural acceptance of personalization, and what Professors Anthony Casey and Anthony Niblett termed “microdirectives” might solve the problem of those traditionally thought of as the dangerous few. Part III will complicate this utopian picture by introducing the idea of the “politically dangerous few”—people who take violent action not because of bloodthirst but because of a rational belief that violence is necessary for their political ends. I argue that even, and given the paradox of tolerance perhaps especially, in an abolitionist world, applying personalized law tools to this group exposes those tools’ potential for stymying political dissent if placed in the wrong, even if well-meaning, hands.
II. Identifying and Constraining the Dangerous Few with Personalized Law
I turn first to the promise of personalized law before discussing its peril in the next section. Personalized law might be able not only to identify the dangerous few,3In this section I will use the term “dangerous few” to mean the “interpersonally dangerous few,” meaning those who commit the sort of violent interpersonal harm that traditionally signifies membership in the dangerous few. but it also creates the possibility for non-carceral solutions that were previously unimaginable.
The first of these possibilities is one that is routine within personalized law. Its predictive power far surpasses anything that a human could imagine. The ability to consider not just a few salient characteristics, but hundreds or even thousands of them, all tied not to the rough heuristics humans use for decision-making but instead to thousands or millions of real-world data points, may well lead to the ability to identify not only who is likely to commit serious criminal harm, but when, what, and how they are likely to do so as well.
And this identifying power may be especially heightened in the case of identifying the dangerous few, as opposed to those who would commit criminal harm more generally. That is because the sort of intense oversight of the people likely to fall into this category may well be inappropriate (or politically infeasible) for the general population. No matter how good algorithmic predictions may get, and despite our history with terrorism-related prosecutions,4“‘You’re trying to get people before they commit a crime,’ said David Cole, the national legal director of the American Civil Liberties Union. ‘And if you’re doing that, you’re going to end up locking up a lot of people who probably would have never ended up committing a real crime. Who knows whether any of these people would have engaged in a terrorist act? I don’t know, you don’t know, the courts don’t know, the government doesn’t know.’” it seems highly unlikely that people would allow a full blown Minority Report-like pre-crime regime to take hold. But if this more extensive oversight is only triggered when someone commits some serious harm and it seems likely they would do so again—stereotypically, this might be the child who takes pleasure in harming animals—it seems justifiable both politically and morally. Indeed, in a personalized law regime, it would be incredibly surprising if the fact that someone previously committed harm was not included in the many variables the law considered.
But beyond merely identifying the dangerous few, personalized law might make possible previously infeasible methods of solving this problem through non-carceral means. Here, abolitionists could harness personalized law’s ability to convey microdirectives—individualized legal rules, or even simply personalized pro-social suggestions, given on a continuous basis. Once someone has been identified as being potentially part of the dangerous few, a personalized law regime might send them ongoing suggestions or commands designed to nudge or force them toward making more pro-social choices. A person could be encouraged to attend community meetings to build connections with their neighbors, to go to therapy or take certain medications, or told to stay away from certain people and places. And all of these suggestions or mandates would be no more obtrusive or noticeable to the outside world than the “breathe” or “time to stand” notifications familiar to Apple Watch owners.
The idea of interventions like these are not particularly novel. Requirements to avoid certain places, to attend treatment, or to engage in community service are common throughout the criminal legal system today. What makes them different in a personalized law regime is that each nudge or requirement could be delivered at the precise moment that it would be most effective for the individual, and the range of possible nudges is nearly endless. Through personalized law, a person could effectively be given a step-by-step guide through life.
To some, or perhaps to most, this might sound dystopian. It admittedly suggests a level of state-mandated control that could exceed all but the most burdensome conditions of a physical prison. Indeed, dystopian art has shown just how dystopian a system like this one could be if misused. Further, it assumes the existence of a device that would convey these real-time messages that a person would have with them at all times. This device, I note, seems eerily similar to the ankle bracelets and other home-confinement devices that anti-carceral activists have objected to. Finally, incarceration is about both physical and mental restrictions. The personalized law regime discussed here may lead to increased physical societal integration, but it appears to do so at the cost of significantly decreased autonomy.
The first objection loses some force when it is made clear that not every personalized law communication would be a mandate. Many, if not most, could instead be nudges; suggestions designed to encourage pro-social behaviors in the hopes that those behaviors become the person’s norm. But beyond this, this most intrusive version of personalized law would apply only to those within the dangerous few, and only after the person has already committed some serious, likely violent, harm. This sort of personalized law regime that is maximized on both the intensive and extensive margins is designed solely to apply to those who, but for it, would need to be incapacitated—indeed, incarcerated—in order to prevent serious harm to other members of society. In this way, the personalized law solution is distinctly less carceral than the alternative. Moreover, because personalized law adjusts as it gains additional information, the various instructions and mandates received by someone identified as being in the dangerous few could become significantly less burdensome as time went on, assuming the person spent more time being a positive (or at least not harmful) member of the community.
The second objection, that this personalized law regime is only possible through a carceral, ankle monitor-like device, ignores that unlike an ankle monitor, the device needed to transmit personalized law instructions would be ubiquitous in a personalized law society. Whether through a smartphone-like device or a wearable like a smartwatch, the stigmatizing effects of ankle monitors would be virtually eliminated because every person, dangerous few or not, would have one. Moreover, because personalized law would depend not only on data collected from these individual devices but also data from a web of public surveillance tools like those that already exist in many cities, those identified as the dangerous few would need to worry less about the false alerts that ankle monitors cause if they, for example, lose a GPS signal, or if a person simply forgets the device at home. Personalized law could also prevent this problem, as well as others associated with e-carceration like the aggressive collection of fees for the device’s use, through its ability to improve through iteration. As an algorithm learned which combination of factors suggested an innocent mistake versus those factors that suggested someone attempting to abscond, it would take those past data points into account when responding to future scenarios. Likewise, by assigning the calculation of personalized fees to an algorithm, fees may well be eliminated for the many people driven into debt by their expense, and the aggressive and sometimes shady tactics that some private GPS monitor companies have resorted to in order to collect could be ended.
Finally, the third objection, that a personalized law regime allows greater physical freedom at the cost of mental restraint, is ultimately a question of implementation. Presumably, personalized law would attempt to maximize the amount of both physical and mental autonomy someone within the dangerous few could have while still preventing them from harming others. Admittedly, different people may have different preferences along these lines, but this too could be personalized. While one person might value greater physical autonomy, another might value more fulsome access to information. The idealized personalized law regime discussed in this section could take account of these preferences by, for example, giving the first person freedom to explore their city at the cost of keeping them away from certain online communities, or doing the opposite for the second person who valued that sort of online engagement over the ability to physically interact with others. Nevertheless, this is a balance, and it is one society must remain cognizant of. It does the world little good if, with the goal of maximizing physical societal integration, it constrains the mind of the writer of the next “Letter from Birmingham Jail.”
Personalized law, then, seems to hold significant promise for an abolitionist future. It suggests a way for even the “dangerous few” to continue to be integrated within the rest of society. By using the predictive power of personalized law to identify this group and then maximizing personalized law along both the intensive and extensive dimensions for them, even the dangerous few may not need to be incarcerated. Indeed, to an outside observer in an abolitionist, personalized law world, the dangerous few would appear just like everyone else; checking their device to see which personalized legal rule or pro-social suggestion applied to them at any given moment.
III. Personalized Law and the Politically Dangerous Few
Thus far, I have painted an admittedly rosy, even utopian, vision of how personalized law might be used not only to identify the dangerous few, but how it might also allow carceral abolitionists to maintain their commitment to non-incarceratory solutions and community integration for even those thought to be the “worst of the worst.” Indeed, the previous section largely assumed away the many practical barriers to personalized law that Ben-Shahar and Porat identify in favor of an idealized version of a personalized law regime. On the path to reaching that utopian state, there would obviously be hurdles like data collection, de-biasing, and many years of testing necessary to implement a sufficiently accurate algorithm (or algorithmic system) that could achieve the identification and engagement purposes discussed. But even assuming away such challenges, the use of personalized law to solve the problem of the dangerous few becomes significantly more complicated if we shift our view from interpersonal harm to societal harm.
Discussions of the dangerous few routinely either implicitly or explicitly assume that the people within this category are those who commit grievous interpersonal harm. The dangerous few commit sexual abuse, child abuse, murder, and assault. And they do so repeatedly and unrepentantly. But in truth, it is not clear this assumption is warranted. Even within our current hierarchical and carceral society, there are few people who fit this description. Indeed, most people—even those who commit heinous crimes in their youth—seem to “age out” of their criminal behaviors in their twenties. And the likelihood of continuing to commit crimes, even violent crimes, drops as age continues to increase. The idea that in an abolitionist world—a world with extensive mental and physical health care; where basic needs are uniformly met; with strong and supportive communities; and without the sorts of racial, gendered, economic, and other unjustified hierarchies that are endemic to our current society—in that world, the idea that people may cease committing the interpersonal harm that signifies belonging to the dangerous few does not seem farfetched. That is especially so given the potential for personalized law to both identify these interpersonal dangerous few and to nudge them towards more socially positive behavior. As such, it seems possible that the combination of an abolitionist world and personalized law might effectively eliminate the category.
Instead, it seems far more likely that the people who would continuously commit the sort of violent harm that typifies belonging to the dangerous few would not be irrational or bloodthirsty. Instead, they would be rationally motivated by their political beliefs; specifically, a rational belief that violence, or even simply non-violent but extreme means outside of the accepted political process, was necessary to achieve their political goals.
Someone unfamiliar with abolitionist thought may be surprised that this sort of political violence would remain in an abolitionist future. But, abolitionist praxis, at its best, is ultimately practical even though it is simultaneously wildly aspirational. #8toAbolition, for example, includes both eight broad social changes—demands like defunding the police and investing in care, not cops—and it lists present-day models for how these points have already been put into action. For present purposes, the relevant piece of this combination of the practical and aspirational in abolitionist praxis means that abolitionists virtually never suggest that people will be perfect and harm will be entirely eliminated in an abolitionist future. Instead, “[a]bolition is about how we respond to harm caused and how we respond when we cause harm.” Abolitionists imagine a future where harm is reduced to the extent possible through the strengthening of social services and community bonds and where the societal response to harm is non-carceral when it occurs. But the idea of the dangerous few threatens this non-carceral orientation because it suggests that carcerality is, even if minimized, a necessity. As Professors Nicolas Carrier and Justin Piché once wrote:
[T]he irresolution of the problem of the ‘dangerous few’ appears to transform abolitionism into a de facto minimalist posture. Why even stick to the massively unknown and/or misunderstood abolitionist identity if what is at stake is to reaffirm that the principle of subsidiarity and the critique of heteronomy . . . are valid in a vast majority of currently criminalizable events?
I argue this necessity for carcerality is heightened when a person is willing to use violent means not to cause interpersonal harm but to attempt to undermine the abolitionist society itself. This is a variation of what Karl Popper labeled the paradoxes of freedom and tolerance.5Karl Popper, The Open Society and Its Enemies 226 n.4 (1st ed. 1947). To paraphrase Popper, if an abolitionist society is unwilling to incarcerate even those who are determined to destroy it, that abolitionist society must soon cease to exist.
Especially if we are willing to put aside the issue of subversive or extremist speech,6See, e.g., Michel Rosenfeld, Extremist Speech and the Paradox of Tolerance, 100 Harv. L. Rev. 1457 (1987). Popper’s solution to these paradoxes seems directly applicable. A free society must have some restraints on the bully so that they do not dominate the meek, and a tolerant society must not tolerate intolerance lest “the tolerant will be destroyed, and tolerance with them.”7Popper, supra note 5, at 226 n.4 (1st ed. 1947). Or as John Rawls once wrote: “Justice does not require that men must stand idly by while others destroy the basis of their existence.” Likewise, an abolitionist society must allow some carcerality for those who would undermine its abolitionist nature. But these general statements do little to help us draw distinctions between various political camps, each of which is likely to, rightly or wrongly, paint the other as the intolerant one.
It is here that personalized law, and particularly personalized law in an abolitionist world, becomes distinctly dystopian. Personalized law is a tool that lets us finely tailor and target the goals of the law. But as Ben-Shahar and Porat note, it does little to help us choose those underlying goals. In other words, “personalization technology, for all its promise, cannot provide [the law’s] objective.” Those goals are still ultimately political judgments. What personalized law adds is precision and efficiency. If the underlying goals of the law are just and good, then personalized law is highly beneficial. But if those underlying goals are corrupt, malicious, or merely misguided, then personalized law can be a powerful tool for ill.
This is an especially important distinction to recognize in the abolitionist context because the paradox of tolerance is heightened in an abolitionist world. Presumably, part of the reason that our current society is not abolitionist is that it is incredibly tolerant of intolerance. Both as a social and a legal matter, we have decided in our current world that large amounts of intolerant speech and action are tolerable. Hate speech is legal, and in some circles it is prized. And discriminatory acts are widely tolerated, so long as there is an insufficient paper trail to prove that they were done for the explicit purpose of discrimination. This tolerance of intolerance allows for the proliferation of inequitable hierarchies that abolitionists oppose and that hinder the creation of the egalitarian world that abolitionists imagine. An abolitionist world, then, must be more intolerant of intolerance than our current one. In order to maintain a less inequitably hierarchical society than currently exists, abolitionists will almost certainly need to be willing to take steps to tamp down potentially intolerant acts that we currently allow to proliferate. In turn, this suggests that a broader swath of political action might undermine the maintenance of an abolitionist world than might undermine that of our current world.
The combination of personalized law and a broadened desire to refuse to tolerate intolerance could be a recipe for disaster. Turning back to the idea of the politically dangerous few will illuminate this point. Being labeled part of the politically dangerous few would likely require three overarching characteristics: First, one would need to espouse political beliefs that would undermine or overturn the prevailing societal status quo. Second, one would need to view violence, or similar disruptive action outside of the established political process, as justified to achieve these political ends. Third, this person would need to remain willing to take this politically motivated action despite any nudges, community interventions, or punishments that abolitionist legal and social systems might impose.
As in the context of the dangerous few generally, personalized law would aid in identifying these people by consulting a large swath of potentially relevant characteristics. Indeed, a personalized law regime would likely do a better job of identifying them than our current law enforcement regime. That regime has given rise to numerous misfires in which people seem functionally, even if not legally, entrapped, and has failed to identify numerous people who would go on to commit horrific ideologically motivated violence.
But even if personalized law perfectly identifies every person who would use violent means for their political ends, is that a world we want to create? I posit that the answer to that question is no. Even if we are strongly averse to political violence as a general matter, there are too many examples throughout history of violence being put to noble political ends. Imagine if personalized law had been available to stop Nat Turner, or John Brown, or any of the various Founders of this country who would go to war with England. That these sorts of people would have been branded the politically dangerous few is obvious. Both Nat Turner and John Brown, for example, were permanently incapacitated, i.e. killed, when they did not achieve their violent, but inherently political, objectives. Personalized law in this space, then, does not necessarily enable the pursuit of noble political or abolitionist goals. Instead, it would seem to more effectively empower those who wish to uphold the status quo.
This status-quo-maintaining function of personalized law may seem initially appealing in the context of an abolitionist future. But remember that even in an abolitionist future, people are not perfect. They may still thirst for power and attempt to use that power toward negative ends.
The risk of this negative use of power seems magnified in the abolitionist political context because of the abolitionist’s need for broader intolerance of intolerance. A devious or misguided leader could paint any number of their opponent’s political proposals as an attempt at unacceptable intolerance. They might say, for example, that an educational proposal with a disparate impact was in fact an attempt to recreate now-abandoned hierarchies, and a willingness to exercise the right to bear arms signified a willingness to use violence. While this example may seem farfetched, it is essentially the United States’ response to the Black Panther Party’s Free Breakfast for School Children program.
Because personalized law is a tool to effectuate the purposes of the law, but not to determine those purposes, it contains no inherent backstop to prevent this sort of misuse. Instead, it could be used to target political enemies with ruthless efficiency and accuracy, in a way that is nearly unimaginable today.
This problem of targeting is exemplified through one further complication that the idea of the politically dangerous few highlights for abolitionists and personalized law, and especially for the combination of the two. That complication is the intensity of a belief, and especially a political belief, leading to negative legal consequences. This possibility, and the effectuation of it, is admittedly nothing new. The Angola 3 were infamously held in solitary confinement for decades. In large part, this was because of their continued organizing on behalf of the Black Panther Party even after their imprisonment, and indeed organizing within prison is generally cause for negative repercussions. The FBI’s shift from surveilling Rev. Dr. Martin Luther King, Jr. to sending a letter suggesting that he commit suicide as he became more prominent is yet another ignominious example. But these attempts at quelling political dissent were like firing at a clay pigeon with a blunderbuss. Personalized law would be like using a sniper rifle that fired dynamite as its ammunition—both infinitely more precise and far more destructive.
As Ben-Shahar and Porat explain, personalized law is optimally effective when it operates on both the intensive and extensive margins. But here, that would mean that those identified as the politically dangerous few could be especially targeted across a wide number of dimensions. They could be discouraged or prevented from engaging in a vast number of activities in an effort to quell their political allegiances, and they might have numerous rights or privileges restricted in the name of preventing them from committing political violence. While a prohibition on owning firearms comes most obviously to mind, the use of vehicles in terrorist attacks suggests that they may rationally be prohibited from driving a car. But beyond this, because personalized law would operate across areas of the law, they might also have their lives or work devalued in civil lawsuits, or be given slower speed limits, or less favorable contractual terms, or raised insurance premiums, or any of thousands of other possible disincentives to maintaining their political beliefs. And these restrictions, or indeed just changes to their personalized legal regime, could rationally ratchet up as the intensity of a person’s belief increased.
All of which suggests a final problem that a personalized law regime, even an abolitionist one, may create: the stagnation of political ideas. Fully exploring this problem, and indeed a fulsome exploration of the concept of the politically dangerous few more generally, requires many more words than I have here, and so I flag these issues only briefly.
Political stagnation seems a natural outgrowth of a world in which personalized law allows for the precision targeting of dissidents. Even those people who are in no way the dangerous few would recognize that transgressing the wrong political boundary could significantly negatively affect their personalized law commands. Better then to stay clearly within well-worn political bounds. Ironically, this combination of the heightened paradox of tolerance with personalized law would stymie the very sort of revolutionary political thought that many abolitionists think is necessary to build—or even to imagine—an abolitionist world. More than this, however, it could also hinder the sort of evolutionary political thought that leads us to reevaluate historical figures and ideas once lauded or despised by the status quo.
And so, while its application to the interpersonally dangerous few suggests personalized law’s abolitionist promise, the politically dangerous few suggests personalized law’s abolitionist peril. Personalized law might enable new ways to identify those people comprising the interpersonal dangerous few, and it may mitigate or eliminate the harm caused by them through non-carceral solutions that are otherwise impossible. But because the dangerous few is likely not limited only to those who cause interpersonal harm, but also includes those who would cause societal harm, i.e., who would seek to undo the societal status quo, personalized law must be used carefully. The paradoxes of freedom and tolerance are heightened in a non-carceral society that prizes the removal of unjust hierarchy. Because intolerance must be more fulsomely policed in an abolitionist society, the possibility that political opponents will be tarred with the label of intolerance is heightened. When combined with a regime of personalized law, this possibility is heightened still. As Ben-Shahar and Porat ably note, personalized law does not make the moral and political decisions that undergird the law’s purposes; it simply allows those purposes to be effectuated in a more precise way. And so those in control of the levers of an abolitionist government could target their opponents with laser-like precision while justifying this disparate treatment as necessary to the maintenance of the abolitionist future.
Ultimately, this Essay is not meant either to persuade or to dissuade abolitionists or others who would consider a shift to a personalized law regime from adopting personalized law. As I have discussed, personalized law seems to be an incredibly powerful legal tool, and it may create benefits that are impossible to achieve without it. But, because it still depends on people to make the moral and political judgments that will guide its personalized legal commands, it might also allow bad or misguided actors to target their opponents in ways that would be unimaginable today. Personalized law, then, despite its focus on algorithmic solutions and even in an abolitionist future, may only be as good as the people who implement it.
* * *
Adam Davidson is a Harry A. Bigelow Teaching Fellow and Lecturer in Law at the University of Chicago Law School. He thanks Omri Ben-Shahar, Ariel Porat, and the participants in the Personalized Law Symposium for their discussion and suggestions. He also thanks Aneil Kovvali and Elizabeth Reese for their suggestions on an earlier draft and the University of Chicago Law Review Online editors for their work on the piece.
* * *