Imagining Perfect Surveillance


How would society react to “the Watcher,” a technology capable of efficiently, unerringly, and immediately reporting the perpetrator of virtually every crime?  This Essay treats that speculative question as an opportunity to explore the relationship between governmental surveillance and criminal justice.  The resulting argument is unabashedly fictional but draws attention to pressures that may influence the real world.  For instance, the Watcher casts doubt on perfect surveillance’s ability to improve the law, supports judicial attentiveness to substantive law when reviewing rules of investigation, and suggests that legislative control might displace prosecutorial discretion.  The Watcher also draws attention to the relationship between surveillance and regulatory intricacy, as well as to ways of preserving human mercy within automated criminal justice.


Introduction: Why Imagine Perfect Surveillance?

In recent years, three interconnected trends in digital technology have been transforming the law and practice of governmental surveillance.  First, data aggregation has enabled, for example, the National Security Agency (NSA) to monitor much of the data traffic linked to the United States, as well as other countries.1  Second, the smartphone revolution and the “internet of things” have generated unprecedented kinds and amounts of individualized data streams ranging from locational data to social media posts to heart rate information.2  Finally, various forms of digital automation, including artificial intelligence, are increasingly supplanting human judgment.  Examples include automated red light and speed limit cameras, as well as new forms of predictive policing, whereby computers help direct police to sites of anticipated crime.3  These mutually reinforcing trends have made it possible to believe that “perfect surveillance,” or an unlimited ability to detect crime, may lie in our future.4

This Symposium is an exercise in forecasting: The participants have been asked to imagine an important technological development that will challenge and disrupt the law in 2030.  There are many possible ways to approach that daunting task, and one is to engage in the imaginative exercise of fiction.5  Instead of starting with the present and trying to build forward year by year, we could jump ahead to a time when existing trends in digital surveillance have been actualized to an extreme degree.  And instead of limiting ourselves to abstractions or exploring innumerable possibility branches on an elaborate decision tree, we could pursue just one possible sequence of consequences in a rich and detailed way.6  The specific results would be, well, fictional.  But they may also be instructive.  The outcomes that are predicted in speculative fiction often have a way of revealing the pressures that actually influence the real world.

What would the world look like if three emerging features of governmental surveillance—data aggregation, individualized data streams, and digital automation—were realized to a perfect degree?  The short answer is that the government would possess what this Essay imagines as “the Watcher,” that is, a technology capable of efficiently, unerringly, and immediately reporting the perpetrator of virtually every crime.  Such a technology would obviously pose many quandaries for the law.  Three broad questions stand out:

  • Would the technological capacity to create a machine of perfect surveillance lead to its actual use and, if so, how would the law limit the machine’s operation?
  • Would surveillance perfection lead to revisions in the criminal law, such as to decriminalize certain activities or to reduce the harshness of existing penalties, as some scholarly commentators have suggested?7
  • How would the relative powers and responsibilities of lawmakers, courts, and other institutional actors change in light of perfect surveillance?

These questions implicitly hold constant all variables other than the Watcher, even though the radical changes necessary to make the Watcher a reality would certainly bring many other disruptive events.  So, again, the results will be fictive.

Still, isolating the Watcher’s potential effects can reveal the direction in which emerging features of governmental surveillance will tend to push the law.  In other words, a fictional exercise can operate as a thought experiment and so yield insights about where we are today and where we are headed.

Confirming as much, the prospect of perfect or near-perfect surveillance is already a recurrent trope in academic and public debates.8  Sometimes, perfect surveillance is presented as an aspirational ideal.  At other times, it’s viewed as a conversation-stopping disaster.  The most persuasive view is more complex than either extreme.

Without further ado, let me transition to a different speaker at a different symposium—whose views, needless to say, are not necessarily my own.9

I.  Keynote Speech of John Watson, Ph.D.10

Being a partial transcript of the proceedings at the inaugural “Watchathon,” an interdisciplinary teleconference on surveillance held at the Virtual University of America on April 15, 2031.

Remarks by John Watson, Ph.D., Keynote Speaker

When we first launched the Watcher back in 2026, we were trying to realize a dream: a machine that could efficiently, unerringly, and immediately report the perpetrator of virtually any and every crime.  There were a lot of different opinions about what would happen on the way toward that achievement.  Now, five years later, we finally seem to be on the brink of making perfect surveillance a global reality.11

Still, the path to surveillance perfection hasn’t been exactly what we expected.  For example, many of us thought that perfect surveillance would foster decriminalization or quickly lead to automated criminal justice—but neither of those predictions has really panned out.  At the same time, almost nobody predicted that the Fourth Amendment would become, in effect, a means of reviewing substantive criminal laws, or that prosecutorial discretion would dwindle while legislatures and juries enjoyed unprecedented power.

Perhaps most importantly, we’ve come to realize that our attitudes toward crime, investigation, and regulation all depend on the prevailing conditions of surveillance.

My remarks today will provide a retrospective on the Watcher’s first five years of operation, from the perspective of an insider who has been at the vanguard of developing perfect surveillance.  Along the way, I’ll also try to point out all the various predictions that have ultimately been proven wrong—including some of my own.

A.  Turning on the Lights

By the time the Watcher received final legislative approval in 2026, the union of surveillance and artificial intelligence was already a deeply ingrained feature of everyday life.  We had long ago gotten used to knowing that our every act and word was logged and recorded for later analysis.  We knew that our TVs habitually listened in on our conversations, that our watches would offer timely driving tips, and that our refrigerators had opinions on what to cook each night.  And not just our own devices, but also the devices on nearby persons, drones, and structures would inevitably note us as we went about our day.  So from one standpoint, the Watcher was just another pair of digital eyes amidst a thousand other eager onlookers.

But the Watcher was still revolutionary.  Unlike the NSA programs and private data merchants that preceded it, the Watcher mined all existing data sources, including almost all digitally recorded information on networks and the cloud, as well as data streams from special government sensors that private companies couldn’t use.  The Watcher alone had legal authority to analyze all that information with unprecedented efficiency, using a series of algorithms made possible by a decade full of technological and analytical breakthroughs.  A programmer by trade, I had contributed to the design of those algorithms in the confident belief that the Watcher would change the world for the better.

People sometimes compare the Watcher to Sherlock Holmes, that most efficacious of fictional detectives.  But, perhaps because of my name, I always felt that Holmes and all the other super sleuths were overrated.  No matter how brilliant he was or how hard he worked, a flesh and blood Sherlock couldn’t dent the overall crime rate or deliver a meaningful dose of justice to society at large.  But the Watcher could.  Instead of picking up on obscure clues and doing logical summersaults to deduce its way toward improbable conclusions, the Watcher simply did its best to see everything—and it was smart enough to connect the dots.

At first, there was some resistance to the idea of bringing the Watcher online at all.  While it may seem odd in retrospect, the Watcher made many people think of George Orwell’s dystopian vision in his novel 1984,12 which millions of secondary school students still read every year.  But the bogeyman of Big Brother represents tyrannical politics.  And it was easy—too easy, really—to imagine uses for perfect surveillance that were not only un-tyrannical, but positively democratic and liberating.13

The most obvious examples involved crime control.  Many of the Watcher’s most ardent supporters had lost friends or loved ones to terrorism or murder.  And many others were themselves victims of sexual assault, or knew people who were.  Those constituencies alone were enough to earn the Watcher Project near-majority support.  And now, of course, we know that perfect surveillance can be used to address many other forms of oppressive behavior that private persons would otherwise perpetrate against one another.

So we had come to see that Orwell’s twentieth-century worldview had it backwards.  Before the Watcher, the balance of power between society and individual actually favored the individual.  Isolated threats of terrorism and crime routinely created a million little tyrannies that were just as awful as anything Big Brother ever did.  As compared with that perverse baseline, the Watcher would be an equalizer.  It would prevent any person, or kind of person, from ever again inflicting terror on society.

Other critics feared that the Watcher would mean “the death of privacy.”14  What they meant, presumably, was that the Watcher would record and analyze virtually all personal information.  But information privacy actually survived and even flourished.  The fact that the Watcher knew something about you didn’t mean that your neighbor would know it, or that any police officer would either.  Privacy laws sharply limited the kinds of information that the Watcher could share with government agencies or private data merchants.  And those restrictions were written deep in the Watcher’s code, making them almost unbreakable.

In a similar spirit, many worried that the Watcher might be abused by its operators or hacked by third parties.  Fearing that a more powerful government could not be kept within lawful bounds, some critics hoped to prophylactically disempower the state by derailing the Watcher.15  The logic rested on an insistent defeatism: Since the Watcher’s tremendous surveillance power couldn’t be safely contained, it shouldn’t be allowed to exist at all.

These worries were overstated.  In fact, it was always hard to access the Watcher.  Its database and software were protected by multiple layers of classified encryption.  And, as a matter of hardwired code as well as formal law, even authorized personnel could glean only limited information—namely, information about recently committed crimes.  So the Watcher would never return information simply to satisfy an analyst’s voyeurism or some police officer’s idle curiosity.  Anyway, the Watcher was largely self-securing: Any effort to tamper with the Watcher’s circuitry or otherwise exploit its abilities would itself be a criminal act—and so would appear in the Watcher’s reports to law enforcement.  These checks greatly limited the risk of governmental abuse.

But ensuring that the Watcher complied with the law was never enough.  We also had to maintain public confidence in the efficacy of perfect surveillance.  So the Watcher was allowed to operate only when certified to be in good working order after undergoing various tests, like detecting scheduled actions designed to resemble crimes.  In addition, there were periodic audits of the Watcher’s reasoning and the evidence it adduced in reports.  Perhaps most importantly, the Watcher posted all of its search results online within seventy-two hours—with a few exceptions, such as to suppress the names of sexual assault victims or to avoid tipping off criminals who didn’t realize that they’d committed a minor crime on the way toward committing a larger, preventable one.  These disclosure duties were critical to maintaining confidence in the Watcher as it took on an increasingly important social role.  Just imagine the panic that would ensue if the public grew to depend on the Watcher—but then lost faith in it.

Early on, some surveillance moderates argued that the Watcher should be asked to detect only especially serious categories of crime.  And initially, that point of view succeeded.  When the Watcher first went online, it was directed to observe only certain violent felonies, like murder and rape, as well as threats relating to terrorism and national security.  The result was a bifurcated investigative process: one track marked by an automated perfect surveillance system and another track resembling what police had been doing for generations.16

But that didn’t last long.  As though on cue, a financial fraud conspiracy broke just a week after the Watcher was brought online, and the culprits managed to flee the United States with most of the money.  People understandably wondered whether the Watcher could have prevented a few Wall Street crooks from walking away with thousands of average people’s retirement savings.  And the answer, of course, was “Yes.”  So the Watcher was promptly asked to detect and report all felonies punishable by at least ten years of imprisonment.

But that, too, did not last.  People accused of misdemeanors and low-level felonies started asking whether it was possible to retroactively query the Watcher to find out whether they could be exonerated.  And, again, the answer was “Yes.”  Due process, as well as basic common sense, seemed to support using the Watcher to exonerate the innocent.17  In a matter of months, the Watcher was fielding thousands of queries involving low-level crimes.  Eventually, experts and the public alike came to terms with a new reality: Most traditional law enforcement efforts had already been rendered superfluous.  So, over time, the Watcher gained the authority—and the responsibility—to report on almost every kind of criminal offense.

Another moderate proposal was popular among intellectuals—including, I might add, some attendees at this conference.  The basic idea was that a limited amount of criminality is worth having, even celebrating, because some activities that are currently illegal shouldn’t be.  The most trumpeted examples were prohibition and sodomy—two crimes that were once widely enforced but have long been abandoned as senseless and wrong.18  With the benefit of hindsight, the government’s previous surveillance shortcomings began to be viewed in a positive light.  Imperfections in traditional surveillance had created room to break the law, thereby mitigating the law’s unjust consequences while allowing the public to glimpse how the world would be if the law in question were ever repealed.19

The problem was how to put any of this into practice.  Sure, some criminal regimes seem like mistakes in retrospect.  But it’s in the nature of mistakes that you don’t know when you’re making them.  To overcome that problem, some suggested that the Watcher could be limited to detecting no more than, say, one out of every three crimes.  But imagine explaining that lottery to victims.  Others proposed limiting the Watcher to roughly the same crimes that would have been detectable via traditional methods, thereby perpetuating roughly the same level of investigative power over time.20  But wasn’t the whole point of the Watcher to improve on our ability to catch bad guys?

B.  Detection and Decriminalization

Lots of people thought the Watcher would foster a measure of decriminalization or other legal improvements21—and I confess that I was one of them.  But we were wrong.

The decriminalization theory went something like this.  In the past, many crimes seem to have remained on the books only because they were so rarely or selectively enforced.  The narcotics laws were the prime example: Most drug laws wouldn’t have lasted a week if they were enforced against the children of prosecutors, legislators, and CEOs.22  Yet the laws persisted because they were enforced disproportionately against politically vulnerable persons and groups.

The Watcher was expected to improve this situation in two basic ways, one relating to the police’s abilities and the other to their incentives.  First, the Watcher would hand police all the evidence they needed to bring charges against people who could otherwise hide behind high walls and white-shoe lawyers.  So for the first time, it would be equally efficient and practicable to enforce the law in gated communities and on street corners.  With the Watcher, all people would stand as equals, whether rich or poor, powerful or weak.

Second, the Watcher would give police an incentive to engage in more evenhanded enforcement.  As noted earlier, the Watcher generally posted its results online within seventy-two hours of detection.  This meant the public had nearly instant access to the Who, What, When, Where, and How of every reported offense.  So if police selectively chose not to arrest and prosecute criminals, then people would immediately find out about it.  Armed with this new information, the public would presumably demand either that the criminal laws be equally enforced or that they be repealed.

In short, the Watcher would put everyone in the same highly monitored boat.  And when faced with the choice to sink together or swim toward decriminalization, people would go with the second option.  Or so the popular theory went.

Things actually played out rather differently.  Consistent with the decriminalization theory, the Watcher did indeed demonstrate just how many people were committing crimes, including narcotics offenses.  The results overwhelmed the criminal justice system.  In just its first day of operation, the Watcher reported far too many offenders to arrest, even if the police had wanted to.  And there wasn’t any obvious path toward full enforcement.  The costs of incarceration alone were prohibitive, leading many people, both in and out of the government, to protest overly harsh criminal laws.  So far, so good for the decriminalization theory.

Meanwhile, however, thousands upon thousands of criminals were being publicly identified.  Though the law couldn’t immediately reach most of those perpetrators, social sanctions could.  And, having been stigmatized as offenders, perpetrators found that they couldn’t effectively lobby for decriminalization.  Even relatively obscure crimes that previously lacked any serious stigma suddenly became more salient and seemingly more offensive.  And with prosecutors having to exercise so much discretion over when to press charges,23 many wealthy or politically connected offenders focused on obtaining mercy from the executive rather than changing the underlying laws.

At the same time, many employers and parents turned out to like the prospect of finding out about workers or children who were using drugs or otherwise engaging in undesirable conduct.  And as the scale of various crime epidemics became increasingly obvious, public alarm hardened some lawmakers’ resolve.24  It’s one thing to know there’s a lot of crime out there; it’s another thing to see daily lists of offenders, backed by proof of their guilt.  More generally, the publicity that followed the Watcher’s gaze was perfect for sensationalist media, which made a fortune fostering a popular turn toward increased punitiveness.

For example, the rebooted program America’s Most Wanted no longer focused on open cases.  Instead, it reviewed recently identified criminals and argued that they should be arrested, convicted, and punished, with people voting their opinions from home.  The ratings were off the charts.  The Watcher thus enabled large swaths of the public to clamor for increased law enforcement, including on an offender-by-offender basis.  Needless to say, these efforts often reflected various prejudices.  So the Watcher did end up changing attitudes toward crime—it just wasn’t the uniform turn toward mercifulness that so many of us had expected.

With powerful social forces both for and against decriminalization, some new development had to break the tie.  And then something did: The incidence of Watcher-monitored crimes began to decline.  As this trend toward a different kind of “decriminalization” accelerated, it became feasible to incarcerate or otherwise punish a larger fraction of detected offenders.

Changes in the crime rates didn’t happen overnight, of course.  With public opinion divided on whether to loosen criminal punishments, the authorities muddled through, exercising tremendous discretion in identifying the most serious—or least sympathetic—perpetrators.  Meanwhile, all detected offenders knew that the Watcher had identified them, as their names and crimes had been posted on a publicly accessible website.  Now that public identification had been guaranteed and the odds of arrest were greatly increased, many offenders found a way to stop offending.  The trend was most pronounced for violent crimes, which police vigorously enforced.  But it was also visible for narcotics crimes, even though police initially made drug arrests for only a small portion of detected offenders.

As crime rates fell, compromise became possible.  The federal government led the way.  On the one hand, the federal government declined to seek prison time for most narcotics offenders detected in the first two years of the Watcher’s operation and permanently allocated vastly more funds for drug treatment and rehabilitation programs.  But, on the other hand, most criminal penalties remained the same or even higher, so as to maintain the Watcher’s prospective deterrent effect.  Similar deals followed in the states.  This effort at transition basically panned out: Narcotics offense rates continued to decline, and law enforcement was able to act on an increasingly large fraction of the Watcher’s reports.  Before long, the already limited popular movement to radically scale back the narcotics laws had run out of steam.

With each passing month, legislatures were asking the Watcher to monitor additional offenses, and the same story unfolded.  Initially, detection outstripped the will and capacity for enforcement, yielding a discretion-filled period of muddling through.  True, some offenses were scrapped or scaled back due to public opposition, and outcomes varied by jurisdiction.  But on balance, a familiar law-and-order politics triumphed, leaving preexisting criminal prohibitions roughly intact—with penalties about as harsh as before.

C.  The Regulatory Renaissance

Most of that happened under Version 1.0 of the Watcher, which was designed to detect comparatively easily observed physical offenses like murder, theft, and drug possession.  These crimes were tractable because they had relatively objective definitions and tended to occur at discrete places and times.  Further, their mens rea requirements often played out in directly observable ways: The Watcher usually had little trouble seeing when people knowingly used narcotics, talked about planning a theft, intentionally took aim with a weapon, and so forth.

To conserve computational ability—and, admittedly, to placate critics—we initially directed the Watcher to issue reports only when it estimated a greater than 99 percent likelihood that an identifiable crime had taken place.  In other words, the Watcher initially reported only relatively clear-cut violations.  And based on rigorous external testing conducted by independent experts, we found that the Watcher’s false positive rate consistently remained below 1 percent, just as it was supposed to.25  The false negative rate is harder to glean, but we estimate that, for many categories of crime, it dropped below 10 percent.  In practice, when the Watcher identified someone as a criminal, human experts agreed about 99 percent of the time; and the Watcher failed to observe as few as 10 percent of actual offenders. Before long, the public started trusting the Watcher’s judgements more than the experts’.

Still, there are important caveats behind those numbers.  Most importantly, the impressive figures don’t reflect the many crimes that were so ambiguously defined as to stretch or defeat the Watcher’s analytic powers, thereby preventing detection.  Unsurprisingly, certain mens rea questions proved to be especially difficult to define and, therefore, to detect.  To this day, people sometimes point to the Watcher’s early false negative rates for certain crimes as an embarrassment.

But if there was a shortcoming here, it lay more with the vagaries of the law than with the Watcher.  After all, if the Watcher couldn’t reliably ascertain someone’s mental state despite its many eyes and artificial intelligence—which had access to all reported judicial decisions, by the way—then how could you entrust those same judgments to mere human beings?  Often, the Watcher knew even more about ourselves, and our consciences, than we did.

And, to their credit, legislators recognized as much.  To correct the issue, lawmakers set about redefining various crimes so that they would be more easily ascertainable and less imbued with moralistic language.  For instance, a jurisdiction that defined murder as “homicide with a malicious heart” might switch to the relatively anodyne and programmable phrase “homicide with a premeditated intent to kill.”  Some jurisdictions applied a similar approach to sexual assault by specifying what statements or actions would qualify as “consent.”  These more observable offenses greatly increased the Watcher’s scope and accuracy.

In a similar vein, Congress decided to crack down on some groups’ increasingly disruptive efforts to escape the Watcher’s gaze.  The solution was a new crime: watcher impairment, which prohibited any deliberate attempt to impair the Watcher’s operation.  Police enforced the new offense with special enthusiasm, arresting anyone who tampered with the Watcher’s sensors or data feeds.  But most efforts at resistance were pretty amateurish, and the Watcher had so many eyes—both at home and abroad—that it could usually tell when it was being undermined.

Legislators also adjusted to the Watcher by enacting a slew of new “warning sign” crimes, or minor offenses that tended to be committed on the way toward more harmful conduct.  For instance, some states made it nominally illegal to transport otherwise lawful firearm ammunition more than a mile without spending a few moments to register that intention with a web database.  The idea was to prevent mass shootings by allowing the Watcher to report armed individuals who, for malicious purposes, failed to register their movements.  These warning sign offenses brought even more into the Watcher’s view than existing inchoate offenses, such as criminal attempt or conspiracy.

Roughly a year after the initial launch, we rolled out Version 2.0.  With the benefit of more time to study the Watcher’s data stream, we developed the storage capacity and algorithms necessary to detect more complex crimes.  These included offenses that take place not at any discrete time or place, but rather in many incremental phases spanning hours, days, or weeks.  This meant, for example, that the Watcher could trace relatively diffuse criminal conspiracies, which brought a host of additional offenses into view.  Whereas Version 1.0 could have detected a drug courier, for instance, Version 2.0 could identify the lookouts as well.

Even more important, the Watcher was now able to identify the leaders of organized crime, even if they remained far removed from day-to-day criminal operations.  It took years of hard police work to bring down Al Capone, even though everyone knew he was a criminal.  Thanks to the new and improved Watcher, mafia dons and drug kingpins started to be identified with ease—and promptly brought to justice.

The new version also led to a massive crackdown in relatively mundane financial crimes.  My favorite example is tax evasion.  Once the Watcher started keeping a running tab on every individual’s estimated tax liability, most tax filings became superfluous, saving the economy billions of dollars.  There were always little hiccups, particularly given the frequent and abiding ambiguity in the tax laws.  But the Watcher 2.0 was still reliable enough that it more or less made it impossible for people to keep significant income off the books.  The grey and black markets thus came into the light.  The ultimate economic effects are still being debated, but it is fairly safe to say that tax revenues generally went up—way up—as a result.  Naturally, a lot of the resulting revenue gains went toward funding expansions of the Watcher and related enforcement efforts.

Still, some significant forms of crime remained.  For one thing, people sometimes committed crimes without realizing it, or without having a reasonable opportunity to realize it.  While the Watcher reduced these crimes by increasing the incentive to understand one’s legal obligations, mistakes continued to be made.

For another thing, irrational and often unforeseeable crimes can result from mental illness, substance abuse, or a “heat of passion,” and the Watcher would not necessarily deter those behaviors.  Even so, some of these offenses were preceded by inchoate or other warning sign crimes.  For example, an attempt or conspiracy to obtain narcotics may predate a successful, addiction-creating purchase, and the Watcher’s detection of those early offenses would lead to crime-preventing interventions.

More worrisome were offenses that are knowingly self-destructive and insensitive to conventional deterrence, such as suicide bombings.  In general, the more complex the crime, the more opportunities to trigger a warning offense and thereby come within the Watcher’s gaze.  Yet even terrorism may not be preceded by criminal activity, such as when someone lawfully possesses a firearm and then uses it to murderous effect.

The police, too, began to undergo a profound and—in my view—beneficial transformation.  The Watcher avoided many of the tangible harms associated with traditional surveillance, such as broken doors and demeaning pat-downs.  And, with a warrant, the Watcher could be tasked to report an offender’s location throughout any twenty-four-hour period, so there was no longer any need to pay for extensive patrols or widespread searches of persons and property.  There was likewise no great need for interrogations, coerced or otherwise, or for twisting informants’ arms.  In sum, police officers had much less opportunity to be either intrusive or violent.26

On top of that, crimes committed by police were also detected, so enforcement against police officers never again required a crack in the blue wall of silence. Further, civil rights statutes made it a crime to violate clearly established constitutional rules.  So, for example, the Watcher could observe police officers who used constitutionally excessive force.  Not only were the resulting reports used in criminal cases against the police, but they were also used to inform police disciplinary and training practices.  And plaintiffs used the reports to bring civil suits against the officers

But while traditional modes of investigation became obsolete, the police were still left with several critical functions.  Most importantly, police had to detain undeterred offenders for trial, including by force.  That work alone continued to pose many risks and challenges.  Further, police responded to a variety of emergencies that the Watcher didn’t prevent, such as violence stemming from mental illness.  The police also operated preventively by hardening defenses and attempting to anticipate crimes that could be carried out without triggering a warning sign offense.  For example, police kept blocking off streets to protect presidential convoys from being struck by nearby vehicles.

So police still worked to protect and serve, but they were a smaller, sleeker, and more effective force.  Thanks to perfect surveillance, fewer officers were better able to achieve law and order—just as theories of panopticism would predict.27

Around this time, Congress loosened some of the statutory restrictions on how the Watcher could be used.  While many predictive policing strategies were being employed throughout this period, the Watcher was initially asked to look only for completed crimes, including inchoate crimes like attempt.  Sometimes, however, the Watcher saw something to suggest that a crime might suddenly occur, without anyone committing a warning sign offense.  Imagine a spouse who is prone to violence and about to walk in on adultery.  Version 2.0 was able to recognize situations like that.  Where a future crime seemed likely—or, for serious crimes, significantly possible—the Watcher confidentially informed the authorities of the relevant risk.28  This helped to round out the government’s already considerable efforts at predictive policing.

Hoping to do even better, legislators got more creative.  Thanks to declining crime rates, slashed investigative costs, and higher rates of tax collection, the government felt rich.  And the marginal costs of additional enforcement efforts were lower than ever.

Realizing all this, legislators began to enjoy a golden age of regulation—a veritable regulatory renaissance.  Why enact a law crudely prohibiting driving over sixty miles per hour when you can create offenses that are so reticulated that they respond to every additional mile per hour accelerated?  Why define burglary as breaking and entering “in the night time” when you can costlessly establish graduated punishments for sundown, a minute after sundown, and so forth?  When fine-grained detection has virtually no marginal cost, lawmakers can draw some very fine distinctions indeed.29

Other lawmakers went a step further and sought to ban behaviors that seemed unhealthy, antisocial, or impolite.  Instead of waiting for peer pressure to push teenagers to try illegal drugs, some legislators wanted to criminalize hanging out with drug users.  Other legislators proposed to ban all deceptive statements and to more harshly punish deceptive statements to friends, with the most severe sanctions reserved for evasiveness toward one’s parents.  And still others advocated a ban on being late for meetings, as well as a misdemeanor offense for failing to brush one’s teeth twice a day.  And why stop there?

D.  Rights Renewed

Then, finally, the courts stepped in.

Many felt that this development was long overdue, given the Watcher’s unprecedented surveillance activities.  But the Watcher had occupied a gray zone in the law of surveillance.  On the one hand, the Watcher represented a “dragnet” in that troves of private information were accessed, processed, and analyzed as a matter of course, without any prior suspicion.  So the Watcher arguably engaged in investigative searches without individualized suspicion, contrary to longstanding Fourth Amendment precedent.30

On the other hand, the Watcher didn’t report any private information other than identified crimes and supporting evidence—the kind of thing that the government has an interest in knowing.  Moreover, courts appreciated that the Watcher offered many improvements apart from its accuracy.  For example, the Watcher operated strictly pursuant to law, and—unlike many flesh and blood investigators—it eschewed reliance on prejudices or stereotypes to determine whether a crime was taking place.  From the beginning, in fact, the Watcher’s programming prohibited it from fixing its sights on someone simply based on a generalization regarding their race or ethnicity.31  The importance of this change is hard to exaggerate.32

In the end, courts felt that the Watcher’s surveillance benefits outweighed its costs—provided that internal procedures ensured that it wasn’t directed toward nefarious ends.33  So, in a parade of cautious opinions, the judiciary more or less uniformly upheld the Watcher as reasonable under the Fourth Amendment.  That trend hardened into a jurisprudential consensus, as the Watcher became ever more popular.34  Naturally, I found all of this entirely welcome and unsurprising.

But courts were roused to action when legislators’ creativity caught up with the Watcher’s abilities.  The prospect that government might micromanage people’s personal, social, and intellectual lives was simply too startling.  Some of these laws were going nowhere because they violated, for example, First Amendment freedoms of speech and religious exercise.35  That constitutional clutch alone took care of laws prohibiting mere thoughts, political dissent, and religious heterodoxy.  These decisions were legally easy but politically critical; by insisting on adherence to established law, the courts prevented the Watcher from being used to subvert democracy and basic freedoms.

What’s more, courts interpreted the Due Process Clauses to prohibit coercive action that rested entirely on mere predictions of future crimes, as opposed to some already accomplished physical act.36  This meant that predictions alone couldn’t justify arrests, much less criminal convictions.  So despite a mountain of science fiction writing—or, perhaps, because of it—the Watcher couldn’t be used to police future crime.37

Still, many other proposed regulations, like state law bans on forgetting to brush your teeth, seemed at least arguably constitutional even if they pushed the government’s regulatory agenda further than ever before imagined.  But if the Watcher had taught us anything, it was that a profoundly changed surveillance environment can bring about equally profound changes in social attitudes.  So people might eventually have gotten used to the new micro-regulations, much as they had many of the Watcher’s transformative effects.  As it happened, however, the public didn’t get that chance.

In a landmark decision aptly captioned Hyde v. United States, the U.S. Supreme Court allowed many offenders to conceal their criminal activities from the Watcher.  More specifically, Hyde held that Fourth Amendment objections to the Watcher required courts to engage in substantive review of criminal prohibitions.  My lawyer friends tell me that this development was, to say the least, an unexpected fusion of substantive and procedural review.38

Hyde itself involved a previously obscure federal regulation that made it a misdemeanor to claim a certain healthcare tax deduction without consuming a certain amount of vegetables per week during the taxable year.  The tax deduction was supposed to encourage healthy eating, the misdemeanor was supposed to keep people honest, and the Watcher was brought to bear in order to make the deterrent effective.  Simple enough.  Except that the regulation inadvertently dredged up widespread popular anxiety about federal “broccoli mandates,” which had been a bugaboo during the debate over the Affordable Care Act of 2010.39  Back then, nobody focused on how a broccoli mandate could be enforced.  But, with the Watcher around, it suddenly seemed possible that legislatures could comprehensively regulate people’s diets.

The Court began its analysis by holding that any use of the Watcher implicated the Fourth Amendment’s reasonableness requirement.  That conclusion was defensible enough under existing law.  Under traditional Fourth Amendment principles, police have to get a warrant or have reasonable suspicion in order to intrude on individual privacy.  But the Watcher monitored private places and activities as a matter of course.  If the Watcher were forced to adhere to traditional Fourth Amendment rules, it couldn’t achieve its intended goals.

Instead, use of the Watcher would henceforth be limited to only those offenses that the government could defend under what the Court called an “intermediate scrutiny” standard.  Outside the Fourth Amendment context, the Court had long applied various scrutiny tests to ascertain whether a sufficiently important interest justified the government’s intrusion on a constitutional right.  Borrowing from those cases, Hyde required the government to show that the underlying criminal prohibition at issue was—in the Justices’ words—“well-tailored to achieve a substantial governmental interest.”  What this meant wasn’t entirely clear, and still isn’t.  Scrutiny jurisprudence was always somewhat impressionistic, and post-Hyde decisions were virtually impossible to predict.

But the overall theme of the Court’s new Fourth Amendment jurisprudence was clear enough: The Watcher would be constitutionally barred from implementing criminal prohibitions that struck five or more Supreme Court Justices as substantively weird, silly, or invasive.  Only criminal prohibitions that seemed sufficiently important—in a “you know it when you see it” kind of way—would catch the Watcher’s bright light.  Legislative creativity wasn’t eliminated, but it was cabined.

Naturally, the Court’s novel approach spawned enormous public controversy and occupied dozens if not hundreds of symposium papers, each with its own inventive critique or equally inventive rationalization.  (And I see that the present conference has added a few more!)  But if you take the Court at its word, the reasoning behind Hyde was fairly straightforward.

It went something like this.  In the past, surveillance shortcomings had helped to secure an important if unnamed personal freedom—namely, the freedom to engage in personal experimentation.40  For example, the practical difficulty of observing what happens in private homes had fostered the liberty to experiment with consensual activities in those spaces, including everything from political meetings to sexual encounters.41  And when the government took steps to improve its surveillance powers, such as by enlisting agents to break down doors or point thermal scanners at private homes, then the Fourth Amendment—at least in principle—stood in its way.  So a combination of logistical challenges and Fourth Amendment rules had always protected the right not only to think about, but also to engage in, dissident forms of politics, religion, sex, and private life.

The Watcher threatened all of that.  It eliminated logistical constraints and had so far evaded significant constitutional regulation.  By monitoring virtually every space and activity, the Watcher imperiled a key but implicit constitutional value that the Fourth Amendment had always safeguarded: nonconformity.  The Court’s willingness to protect this implicit value was itself something of a surprise, at least to me.  Could unelected federal judges realistically stand up to majoritarian efforts to use the Watcher?

We never really found out.  Many people still feared the prospect of an ever more conformist political order.  And political movements—aided by social media and other digital technologies—gave voice to widespread fears of governmental overreaching.  It turned out that most people were relieved to see courts put on the brakes.  And so Hyde came to define the Watcher’s limits, which—at least so far—remain firmly in place.

E.  Redistributing Discretion

Finally, perhaps the biggest surprise of all: After just a few years, the Watcher had transformed the institutional roles of the legislature, the executive, and the judiciary.  With the exception of an obscure and somewhat whimsical piece of speculative fiction published in 2016,42 nobody saw this coming.

In general, the three branches of government all have the opportunity to influence the enforcement of criminal law, but each does so at distinct procedural junctures.  The result is a kind of conveyor belt, where the legislature (or a rulemaking agency) is located near the front end; the executive (meaning police and prosecutors) is in the middle; and the judiciary (which encompasses both judges and juries) is at the back.

Before the Watcher, investigative discretion tended to migrate toward the middle, which was and remains dominated by police and prosecutors.  The reasons were many and complex,43 but it’s worth noting that surveillance dynamics played an underappreciated role.

In short, the executive was uniquely situated not only to surveil potential criminals but also to evade surveillance by other governmental actors.  Executive officials had almost complete control over who was investigated based on suspicion of wrongdoing.  Sure, a legislature could enact prohibitions and decree certain enforcement goals.  But the job of implementing those goals would fall to executive officials.  Further, it was generally difficult or even impossible for any other institutional actor to monitor executive enforcement decisions, if only because there were so many discretionary choices that most would inevitably evade review.44 Even when trials took place—which was rare—judges and juries typically had only a dim understanding of the discretionary choices that led to charges.

Again, the details of these developments are complex, too much so to canvass them all here.  All in all, however, discretionary power clearly tended to drain away from both ends of the criminal conveyer belt—and to amass in the middle.

As we’ve seen, the Watcher initially generated too many suspects to target and so afforded the executive more discretion than ever.45  Before long, however, the Watcher reversed longstanding trends by distributing power away from police and prosecutors and toward legislatures, as well as judges and juries.  Slide 1 illustrates this new trend.

Slide 1.  Shift of Enforcement Discretion Away From the “Middle”


This shift occurred because the Watcher radically changed the options available to legislatures.  Instead of having to rely on executive officials to detect violations, the Watcher unerringly followed search protocols that were themselves within legislative control.  So there was no longer any need for thousands of police and prosecutors to make difficult resource allocation decisions when investigating crime.  If a legislature wanted a certain type of conduct to generate public accusations of criminal wrongdoing, it could simply enact a statute reflecting that desire and—poof!—the Watcher would make the wish a reality.

Of course, the Watcher’s public accusation of criminality was not the same thing as pressing charges.  But with virtually all law enforcement sources, methods, and statistics publicly known, enforcement decisions became more easily measured, debated, and controlled.  Legislators and the public started to expect—or demand—that police and prosecutors fully enforce the law, whenever the Watcher detected an infraction.  And by this point, the Watcher was being used to identify virtually all crimes.  The discretion-drenched, “muddle through” attitude that prevailed during the Watcher’s early days was slipping away.  Perfection fostered regularity.

Complicating matters even more, prosecutors suddenly started having a hard time explaining why they were the right officials to make enforcement decisions.46  After all, legislators had an equal or greater claim to represent the public.  And there was another, readily available set of institutions to engage in fact-specific mercy: judges and juries.

In a few exceptional cases, executive officials continued to exert prosecutorial discretion, such as by declining to bring charges against detected offenders.  But that kind of discretion suddenly had to occur in full view of the public, and only the rare prosecutor wanted to be the roadblock stopping prosecution of a clearly criminal act.

That was only half of the institutional transformation that the Watcher brought about.  Because it largely eliminated the need to resolve highly contested empirical questions, such as factual disputes regarding guilt, the Watcher presented a stark question: Should trials, sentencing hearings, and the pardon process—traditional staples of the criminal justice system—become a mere formality, rubberstamping a mechanized system of perfect surveillance?

Several considerations generated an answer of “No.”  First, jury trials and sentencing hearings were deeply entrenched institutions, both legally and culturally.  So instead of simply allowing themselves to be swept away, participants in those institutions—including judges, juries, and attorneys—sought a new, or renewed, social role.  And they did so by looking beyond guilt to consider case-specific grounds for mercy that the criminal law hadn’t contemplated.

Second, relatively impartial judges and juries were better suited than zealous prosecutors when it came to identifying grounds for case-specific mercy.  After all, prosecutors often respond to the same law-and-order dynamic that affects legislatures.  In practice, only judges and juries—potentially majority-defying actors—were likely to exercise mercy in cases where the legislature had authorized the Watcher’s work.

Finally, and perhaps most importantly, some members of the public wanted to preserve a central role for human decisionmakers at some stage of the criminal justice process.  So while the public had previously found it acceptable that human police and prosecutors had largely wrested institutional control from judges and juries, there was much stronger public opposition to a truly automated criminal justice system.47

The solution was that judges and juries started to assert themselves.  Since prosecutors were already approaching full enforcement, defendants had little to lose by insisting on going to trial.  As the criminal trial rate rose, it became more common for defense counsel to tell the jury about the sentencing consequences of conviction and to argue for acquittal despite the evidence.  And judges increasingly allowed it.  It turns out that hearing about mandatory minimum penalties sometimes convinces juries to acquit outright, or convict on lesser charges.  As this kind of nullification became routine, legislatures reduced mandatory minimums in an effort to encourage juries to convict.  And that shift in turn expanded opportunities for judicial mercy at sentencing.  So no matter what pro-punishment legislatures did, they encountered the prospect of adjudicative mercy.

A new jurisprudence of mercy began to evolve.  Depending on the severity and nature of the offense, defense counsel could argue for mercy based on the unique circumstances surrounding the crime, the distinctive traits pertaining to the offender, the excessive harshness of the punishment, or even the basic foolishness of the relevant criminal prohibition.

The front and back end changes wrought by perfect surveillance turned out to be mutually reinforcing.  As noted earlier, legislatures had already taken great strides in making criminal prohibitions more objectively definable, so as to allow the not-so-perfect early version of the Watcher to operate more effectively.  That trend crystalized the widespread sense that trials and sentencing had become mechanized formalities, and so strengthened the case for repurposing the trial for a post-Watcher world.  Counterintuitively, the trend toward rigorous objectification at the front end fostered an increase in discretionary moralization at the back end.

As a result, the adjudicatory process essentially came to proceed in two stages.  First, the Watcher adjudicated factual questions in a way that was objective, determinate, and regular.  Second, human beings made overtly moral decisions that were sometimes subjective, indeterminate, and unpredictable.  This felicitous arrangement promoted confidence in the Watcher, served as a backstop to any glitches or shortcomings in the surveillance system, and, just as importantly, fostered public participation in criminal justice.  The big loser was the executive branch, which saw its discretion dwindle.

And not just dwindle, but qualitatively change.  We’ve already seen that police and prosecutors used to prioritize offenses that were easier to investigate, prove, and prosecute.  For example, executive officials might favor enforcement of relatively easy-to-prove narcotics offenses over complex financial crimes that would require lengthy investigations and attract elite defense counsel.  Thanks to the Watcher, however, those kinds of cost-saving shortcuts no longer made sense.  So instead of focusing on the ease of detection or proof, executive officials became more likely to rest their enforcement decisions on the likelihood of ultimately obtaining assent from judges and juries.  In other words, executive discretion increasingly endeavored to anticipate the discretionary mercy shown by impartial adjudicators.

These days, it’s easy to forget how many people believed that the Watcher would spell the total victory of automated criminal justice.  Yet, as we’ve seen, the Watcher actually brought about an increase in discretionary, often case-specific mercy at the back end of the criminal justice process, through an expansion of jury nullification at trial and judicial discretion at sentencing.  The Watcher’s real effect in this area was to automate much of the law’s execution, leaving nonexecutive human decisionmakers with a greater share of influence.

In sum, the once awesome power of prosecutorial discretion has become a shadow of its former self.  In any given case, there’s little room for prosecutors to advance their own policy goals, as they routinely did in days gone by.  Instead, prosecutors now struggle to implement highly specific legislative decisions, while endeavoring to predict trial outcomes coming down the pike.

*          *          *

Our story has reached the present day, roughly five years after the Watcher brought its bright light to the United States.  I know that other speakers today will be addressing the new research and design proposals that now occupy the public’s attention, especially Version 3.0, the possibility of extending the Watcher to most noncriminal legal disputes, and the exciting new United Nations initiative to spread the Watcher worldwide.

Before I stop for questions, let me end on a note of humility.  People often ask me if I have any regrets about my role in creating the Watcher, and I used to say that the answer was a flat “No.”  But having prepared these remarks, I would now offer one caveat: I wish I hadn’t been quite so confident about how things would turn out.

For as long as I can remember, I’ve thought of perfect surveillance as a wondrous ideal and laughed when people invoked it as a serious cause for concern.  But in hindsight, the truth was always somewhere in between.  Before a transformative technology is unleashed, there’s really no way to know all the ways that it will change our lives.  And I do believe that the Watcher will prove to be the most transformative technology of my lifetime.

So, with feelings of both relief and guilt, I can admit that my critics were partly right: Things could easily have turned out differently.

Conclusion: Back to the Present

Thus spoke John Watson, Ph.D.

What do his fictional words suggest?  Put another way, what can we learn from this thought experiment?  Let me list five primary points below, which correspond to the five Parts of Watson’s speech.

First, if perfect surveillance becomes technologically possible, it will probably be put into practice, permanently shifting the balance of power between government and private individuals.  The promise of even near-perfect personal security would simply make perfect surveillance too politically attractive to resist.  However, there would likely be substantial legal and programmed checks on the operation of perfect surveillance, including requirements of public transparency, systematic efforts at validating results, and the use of perfect surveillance itself to monitor the technology’s government operators.  In combination, these developments would challenge constitutional democracy without overwhelming it.

Second, some commentators may be too optimistic in suggesting that perfect surveillance would tend to foster substantive legal improvements or a reduction in the overall harshness of the criminal justice system.  Of course, the results would vary by crime and jurisdiction.  And greater enforcement transparency would indeed facilitate efforts to increase the evenhandedness of law enforcement.  But familiar social forces, including law-and-order politics, the stigmatizing power of crime, and popular alarm at socially disapproved conduct, would all tend to preserve the existing criminal law, including its breadth and severity.  In short, the politics of criminal justice can keep up with surveillance technology.

Third, perfect surveillance would likely inaugurate a “regulatory renaissance” (Watson’s phrase) by creating a host of entirely new regulatory options for lawmakers.  Crimes that had once been on the margins of observability would become easy to detect, making previously unenforceable laws realistically enforceable.  The door would thus be open to new regulations directed at private life, as well as unprecedented regulatory complexity.  At the same time, policing would become far more efficient, making it even easier for the state to enforce conformity with social norms.  Traditional understandings of the nature and purposes of substantive criminal law would consequently come under strain.

Fourth, courts could effectively cabin the regulatory renaissance by exhibiting greater attention to substantive law when reviewing law enforcement.  Many procedural limits on police, such as the warrant requirement, would severely reduce the effectiveness of perfect surveillance; to avoid that result, courts would likely expand exceptions to traditional Fourth Amendment rules.  By contrast, there would be ample room—and need—for courts to vindicate substantive rights like the freedom of speech, as well as substantive values such as nonconformity.  Further, courts might combine substantive and procedural review of law enforcement by limiting the range of substantive crimes that can be investigated using new surveillance technologies.  If the public became alarmed by rapidly expanding surveillance, then courts may be prepared to make this fundamental regulatory shift.

Finally, the legislature’s gain would be the executive’s loss.  Today, perfect surveillance is the exception, such as when a police officer or red light camera can observe every traffic violation at a given intersection.  But basic institutional principles are built for the normal case.  And when lawmakers consistently have the ability and incentive to legislate investigative priorities, executive discretion will tend to decline.  In essence, a critical dimension of traditional executive power would have been replaced by a programmable, automated process.  As we’ve seen, much of the executive’s lost power would shift to the legislature.  Judges and juries might also see an increase in power—if the public demands some surviving avenue for fact-based, human mercy.  Jury nullification in particular could become an increasingly important failsafe.

If you find any of the foregoing outcomes either reassuring or alarming, the lesson here is not that they are inevitable but rather that they are the possible results of foreseeable pressures.  With the benefit of Watson’s fictional hindsight, we are better prepared to spot the trends that we want to foster, as well as the tendencies that we want to avoid.

[1].      See, e.g., Laura K. Donohue, Section 702 and the Collection of International Telephone and Internet Content, 38 [small-caps]Harv. J.L. & Pub. Pol’y [end-small-caps]117, 152 (2015) (“[T]he NSA was collecting hundreds of millions of records, ranging from metadata to content, transiting fiber optics cables between Google and Yahoo data centers [overseas].”).

[2].      See, e.g., Matthew R. Langley, Note, Hide Your Health: Addressing the New Privacy Problem of Consumer Wearables, 103 [small-caps]Geo. L.J.[end-small-caps] 1641, 1644 (2015) (“[W]earable devices can monitor and record physical activity and sensitive health information—such as a user’s heart rate, skin temperature, or respiratory rate—in real time.”).

[3].      See, e.g., Andrew Guthrie Ferguson, Big Data and Predictive Reasonable Suspicion, 163 [small-caps]U. Pa. L. Rev.[end-small-caps] 327, 330 (2015); Elizabeth E. Joh, Discretionless Policing: Technology and the Fourth Amendment, 95 [small-caps]Cal. L. Rev.[end-small-caps] 199, 209 (2007) (discussing discretionless policing in the context of automated traffic enforcement systems).

[4].      See Christina M. Mulligan, Note, Perfect Enforcement of Law: When to Limit and When to Use Technology, 14 [small-caps]Rich. J.L. & Tech.[end-small-caps] 1, 3 (2008) (“A ‘perfect surveillance’ technology would not interfere with the act of violating the law but would detect every instance of its violation.”).  A significant literature discusses the ability to engage in some amount of “perfect surveillance” or some number of “perfect searches,” meaning the completely accurate detection of specifiable types of crime.  See Michael Adler, Note, Cyberspace, General Searches, and Digital Contraband: The Fourth Amendment and the Net-Wide Search, 105 [small-caps]Yale L.J.[end-small-caps] 1093 (1996) (framing perfect searches specifically as “contraband” searches); Jonathan Zittrain, A Few Keystrokes Could Solve the Crime. Would You Press Enter?, [small-caps]Just Security[end-small-caps] (Jan. 12, 2016, 9:05 AM) [].

[5].      See, e.g., Mitchell Travis, Making Space: Law and Science Fiction, 23 [small-caps]Law & Literature[end-small-caps] 241, 242 (2011); see also [small-caps]Jonathan Glover, What Sort of People Should There Be?[end-small-caps] 17 (1984) (“[O]ur values often become clearer when we consider imaginary cases where conflicts can be made sharp.”).

[6].      For a similarly imaginative exploration in a different but in some respects complementary direction, see Jane R. Bambauer, Glass Half Empty, 64 [small-caps]UCLA L. Rev. Disc. [end-small-caps](forthcoming 2016).

[7].      The thought is that substantive legal reform may result from perfect enforcement—which may in turn result from, or depend on, perfect surveillance.  See, e.g., Jane Bambauer, Other Peoples Papers, 94 [small-caps]Tex. L. Rev. [end-small-caps]205, 226–27 (2015) (“The best way to test whether a criminal statute is appropriately defined and conscribed, and that its penalty is fair, is to aim for more evenly distributed detection so that the costs of a law are felt by the elite and politically powerful.  [I]f [a] senator’s daughter has the same chance of getting caught committing a crime as a relative nobody, an irrational law or unjust penalty will be revisited.”); Arnold H. Loewy, The Fourth Amendment as a Device for Protecting the Innocent, 81 [small-caps]Mich. L. Rev.[end-small-caps]1229, 1248 n.86 (1983) (“A final objection . . . is that if we had a device that would detect evidence of crimes whenever it existed, it would possibly precipitate enforcement of laws that we do not really want enforced. . . . If so, the better course of action is for the legislatures and perhaps the courts to rethink the propriety of [those] laws.”); Paul Rosenzweig, Privacy and Counter-Terrorism: The Pervasiveness of Data, 42 [small-caps]Case W. Res. J. Int'l L. [end-small-caps]625, 645 (2010) (“[T]he use of technology to make it more likely (if not certain) that violations of law will be observed—will work powerful effects on the deterrence component of law enforcement. . . . An interesting corollary to the development of new technologies is that they will, inevitably, require either a reduction in punishments across the board or a much better, and narrower, definition of ‘wrongful conduct.’”); Eugene Volokh, Traffic Enforcement Cameras (Mar. 26, 2002), [] (suggesting in connection with traffic camera surveillance: “I think that broader and more evenhanded enforcement will generally (not always, but usually) lead to improvements in the law.  If lots of citizens get pulled over for speeding, and the limit also ends up making everyone else drive too slowly, City Hall will react.”).

[8].      See, e.g., supra notes 4, 6–7 (collecting sources), infra note 34.

[9].      Throughout, footnotes are for the interest of real life readers.

[10].     With apologies to Sir Arthur Conan Doyle.

[11].     See supra note 4 and accompanying text.

[12].     [small-caps]George Orwell, Nineteen Eighty-Four[end-small-caps] (1949).

[13].     For commentators who find virtue in greater surveillance perfection, see Loewy, supra note 7, at 1244–48 (1983); William H. Simon, In Defense of the Panopticon (Columbia Law Sch. Pub. Law & Legal Theory Working Paper Grp., Paper No. 14-412, 2014), [].

[14].     For a good sampling of the relevant literature, see generally David Brin, The Transparent Society (1998); A. Michael Froomkin, The Death of Privacy?, 52 [small-caps]Stan. L. Rev. [end-small-caps]1461 (2000); Paul Ohm, The Fourth Amendment in a World Without Privacy, 81 [small-caps]Miss. L.J.[end-small-caps] 1309 (2012).

[15].     Cf. Daryl J. Levinson, Incapacitating the State, 56 [small-caps]Wm. & Mary L. Rev. [end-small-caps]181 (2014).

[16].     See Jack M. Balkin, The Constitution in the National Surveillance State, 93 [small-caps]Minn. L. Rev.[end-small-caps] 1, 16 (2008) (“Once governments have access to powerful surveillance and data mining technologies, there will be enormous political pressure to use them in everyday law enforcement and for delivery of government services.”).

[17].     For analogous arguments, see Joshua A.T. Fairfield & Erik Luna, Digital Innocence, 99 [small-caps]Cornell L. Rev.[end-small-caps] 981 (2014).

[18].     See, e.g., Ian S. Thompson, ‘Perfect Surveillance,’ Says Edward Snowden, Could Have Snuffed Out the LGBT Movement. He’s Right, [small-caps]ACLU Blog[end-small-caps] (Feb. 27, 2015, 4:35 PM), [].

[19].     See Bryan H. Choi, The Anonymous Internet, 72 [small-caps]Md. L. Rev. [end-small-caps]501, 558–59 (2013); Jerry Kang, Information Privacy in Cyberspace Transactions, 50 [small-caps]Stan. L. Rev.[end-small-caps] 1193, 1219–20 (1998); Paul Ohm, Good Enough Privacy, 2008 [small-caps]U. Chi. Legal F. [end-small-caps]1, 20–22 (2008).

[20].     See, e.g., Orin S. Kerr, An Equilibrium-Adjustment Theory of the Fourth Amendment, 125 [small-caps]Harv. L. Rev.[end-small-caps] 476 (2011).

[21].     See supra note 7 and accompanying text (outlining actual scholarly arguments describing linkages between surveillance, enforcement, and substantive law).

[22].     See Bambauer, supra note 7, at 226–27 (proposing a “senator’s daughter test”).

[23].     See Elizabeth Joh, The New Surveillance Discretion: Automated Suspicion, Big Data, and Policing, 10 [small-caps]Harv. L. & Pol’y Rev. [end-small-caps]15 (2016).

[24].     See Mulligan, supra note 4, at 14 (“While some very unpopular laws may be changed [in a regime of perfect enforcement], politicians may avoid altering controversial laws for fear of losing their own popularity in a public relations mishap.”).

[25].     Remember: This is fiction.

[26].     Cf. Devon W. Carbado, Blue-on-Black Violence: A Provisional Model of Some of the Causes, 104 [small-caps]Geo. L. J. [end-small-caps]1479 (2016).

[27].     Cf. [small-caps]Michel Foucault, Discipline and Punish[end-small-caps] (1975).

[28].     Cf. Person of Interest (CBS 2011) (imagining “the Machine,” an AI that harnesses mass surveillance to predict crime).

[29].     See Anthony J. Casey & Anthony Niblett, The Death of Rules and Standards (Univ. of Chi. Pub. Law Theory Working Paper Grp., Paper No. 550, 2015), [] (discussing “micro-directives,” or highly context-specific commands generated by machines implementing legislative goals).

[30].     [small-caps]U.S. Const. [end-small-caps]amend. IV; Ferguson v. Charleston, 532 U.S. 67, 79 (2001).

[31].     But see [small-caps]Cathy O’Neil, Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy [end-small-caps]25, 146 (2016) (noting that facially race-neutral metrics often incorporate effects of racism).

[32].     Cf. Tracey Maclin, Race and the Fourth Amendment, 51 [small-caps]Vand. L. Rev.[end-small-caps] 333, 339–40 (1998).

[33].     See Camara v. Mun. Court, 387 U.S. 523, 536–37 (1967) (“[T]here can be no ready test for determining reasonableness other than by balancing the need to search [or seize] against the invasion which the search [or seizure] entails.”).

[34].     Cf. Anthony Amsterdam, Perspectives on the Fourth Amendment, 58 [small-caps]Minn. L. Rev.[end-small-caps] 349, 384 (1974) (raising the possibility of the government altering subjective privacy expectations by “announcing half-hourly on television . . . that we were all forthwith placed under comprehensive electronic surveillance”).

[35].     [small-caps]U.S. Const.[end-small-caps] amend. I.

[36].     [small-caps]U.S. Const. [end-small-caps]amends. V, XIV.

[37].     E.g., [small-caps]Minority Report[end-small-caps] (20th Century Fox 2002).

[38].     See William J. Stuntz, O.J. Simpson, Bill Clinton, and the Transsubstantive Fourth Amendment, 114 [small-caps]Harv. L. Rev.[end-small-caps] 842, 847 (2001) (critically discussing Fourth Amendment transsubstantivity).

[39].     See James B. Stewart, How Broccoli Landed on Supreme Court Menu, [small-caps]N.Y. Times[end-small-caps] (June 13, 2012), [].

[40].     See, e.g., Julie E. Cohen, Examined Lives: Informational Privacy and the Subject as Object, 52 [small-caps]Stan. L. Rev. [end-small-caps]1373, 1426 (2000); Neil M. Richards, The Dangers of Surveillance, 126 [small-caps]Harv. L. Rev. [end-small-caps]1934 (2013).

[41].     Cf. William J. Stuntz, The Substantive Origins of Criminal Procedure, 105 [small-caps]Yale L.J. [end-small-caps]393, 413 (1995).

[42].     See Richard M. Re, Imagining Perfect Surveillance, 64 [small-caps]UCLA L. Rev. Disc.[end-small-caps] (2016).

[43].     See generally David A. Sklansky, The Nature and Function of Prosecutorial Power, 117 [small-caps]J. Crim. L. & Criminology[end-small-caps] (forthcoming 2016), [].

[44].     For discussion of early situations where technology made some measure of discretionless enforcement possible, see Joh, supra note 3, at 209.

[45].     See text accompanying supra note 23.

[46].     See Froomkin, supra note 14, at 1470–71 (suggesting that in a world of “perfect law enforcement”—that is, “a world in which no transgression goes undetected and, perhaps, unpunished”—“the reliance on police and prosecutorial discretion on which our legal system is based will come under severe strain”).

[47].     For a thorough discussion of both the promise and the perils of incorporating machines into criminal adjudication, see generally Andrea Roth, Trial by Machine, 104 [small-caps]Geo. L.J.[end-small-caps] 1245 (2016).

About the Author

Assistant Professor, UCLA School of Law. I am greatly indebted to the participants in the 2016 Program on Understanding Law, Science, and Evidence (PULSE) Symposium, who offered insightful comments and whose example inspired the choice to draft this piece partly in the genre of speculative fiction. I am especially indebted to Kristen Eichensehr, Tyler Jones, and Jaci Seelagy. Thanks also to Will Baude, Sam Bray, Terra Laughton, and the editors of Discourse.

By uclalaw