We are living in the age of Big Data—a time of big possibilities but also big problems. With data abundant, the natural urge of government has been to gather and make something of all that information. What could be better than to eliminate crime, terrorism even, with nothing but information? Focus on that hope often has blinded us to the real difficulties and threats posed by this mass data grab.
In assessing the value of data to policing, it is important to pay attention to the countless instances in which the use of databases erroneously causes law enforcement to target a wholly innocent individual; mistakes here can be costly both to the government and the person involved.
Ironically, though, the mass collection of personal data by the government should concern us just as much when the process is error-free and the data used as intended. That is because the whole point of the mass collection of data so common today is to allow the government to build dossiers on as many of us as possible, on the theory that once the accumulated data is analyzed, offenders will pop out. For the system to work, the government must have access to the data of everyone, innocent as well as guilty, in order to search for patterns that point to wrongdoers. What that means in practice is that the government is building vast databases of our personal information. We all are forced to relinquish our privacy and security, and that should be of serious concern.
The endeavor to collect massive amounts of data in an effort to keep us all safe may be a good idea or a bad one, depending on how such data-collection programs are adopted and implemented. The trick is to use the principles of democratic and constitutional policing to maximize the possibilities of databases, while minimizing the threats to our liberties and privacy. That’s essential, but it is not at all what the government is doing at present.
THE DANGERS OF DATABASES
Meet Abe Mashal—an average Midwestern guy. He’s tall, taciturn, a big fellow with a baby face and a goatee. A dog trainer by profession, on April 20, 2010, Mashal was in Chicago’s O’Hare Airport, en route to Spokane to train a customer’s pets. The trip, though, had gotten off to a bad start. When Mashal tried to check in online the night before, he was told he had to do it with an agent at the airport. So he got to O’Hare early—faced with squeezing his large frame into an airplane seat, he was eager to make Southwest Airlines’s first boarding group. The ticket counter agent typed his name into the computer, gave him a “very strange look,” and then disappeared into the back. Minutes later, Mashal was surrounded by some thirty officers from the Chicago Police and Transportation Safety Administration.1
That’s when the counter agent told him he was on the No Fly List.2
Mashal thought this had to be a joke, “some huge mix-up.” Told the FBI wanted to interview him, he agreed, provisionally. “I will go in back to answer your questions, but I only have one rule.” “What’s that?” they asked. “No water boarding!” Everyone tried hard to stay stone-faced, but many couldn’t help tittering.3
He’s like that; Mashal’s a jokester. The events that began in O’Hare Airport that day turned into a huge ordeal for him and his family; in the course of it Mashal wrote an autobiographical account of his life and what happened to him that for a novice author is pretty decent reading. In it he describes in elaborate detail how he pranked and joked his way through school, never taking anything too seriously—the consummate “class clown.” Like the times he’d set open liquid beverage containers spinning down the school staircase, soaking the unlucky with milk or even urine, to great guffaws from his crew. That lasted until he got busted and was given thirty days janitorial duty in lieu of expulsion, working right alongside the very folks whose lives he’d been making miserable. After a few days of hazing, though, they were all best buddies.4
On leaving school, Mashal cleaned up his act. He decided he wanted to go into law enforcement, become an FBI or Secret Service agent. Those jobs are hard to get but the recruiter he spoke to indicated one clear path. So Mashal took it, and to this day it represents to him a singular life accomplishment.5
Mashal’s a former marine. He made it through boot camp, got married, had a kid (the first of four), got a college degree, became an expert shot, marksmanship instructor, and K-9 handler, and ran a supply shop. Then he left the Corps, deciding against law enforcement—he’d had enough bad bosses in the service to know he wanted to work for himself—and became a private dog trainer. With some others he also started a not-for-profit that trains and provides service dogs to veterans who need them. “Always Faithful Service Dogs.” Semper Fi, that’s Abe.6
So there he was, standing in the terminal surrounded by law enforcement “with my Marine Corps luggage bag, and Marine Corps t-shirt being told that I am on the No Fly List.”7
Several months later, having failed to get himself off the list, despite numerous efforts, including filling out the redress forms the government offered, Mashal became part of an ACLU lawsuit against the Department of Homeland Security. The lawsuit challenged the constitutionality of the way the no-fly list is operated. Mashal is one of a group of plaintiffs, several of whom are U.S. veterans.8
Many of the ACLU’s no-fly plaintiffs have something else in common too: they are Muslims, or of Arab descent. Mashal is both. His full name is Ibraheim Mashal, though he’s always just been known as Abe. Half Palestinian Arab, half Italian. First generation on his dad’s side, second on his mom’s. An American of immigrant stock, like so many of us. Mashal’s book is titled No Spy No Fly because, among other indignities, the FBI offered to get him off the list if he’d spy on mosques for them. That apparently has happened a lot, the government using the no-fly list as leverage to recruit spies.9
To this day, Mashal can’t understand how it is he landed on the list. He is not particularly religious. His dad dragged him to religious school at one point when he was growing up, but that didn’t last long. He wonders if it is because, during a short period of tragedy when he sought religious solace, he wrote to an imam asking for advice about raising children in two faiths. (Right after he left the service, his father died, his wife miscarried eight weeks before their second kid was due, and a close friend had a fatal crack-up on a motorcycle.) In his book he wrote that he “felt ignorant for not knowing how bad these watch lists had become and how many innocent people were affected by it.” These “sounded like stories that would come out of a place like communist Russia. Not America.”10
Welcome to Mashal’s world—one that is fast becoming the world of all of us. We live in the time of massive government databases that keep track of much of what we do. Under the long-standing “Secure Flight” program, the TSA checked fliers’ “name, gender and date of birth” against “terrorist watch lists.” In 2013, the TSA announced it was expanding passenger screening, compiling records of property ownership, physical traits, intelligence and law enforcement data, tax IDs, and travel histories and reservations. Our information is handed out to state, local, and even foreign governments, or “private companies for purposes unrelated to security or travel.” Those who find themselves accused (rightly or wrongly) of TSA violations can be reported to private debt collection agencies.11
In this new world of government databases, once you are caught in the machine, it can prove very hard to set things right. The late Senator Edward Kennedy found himself on the no-fly list. After he was stopped several times before boarding or subjected to additional screening, his staff contacted TSA headquarters. It took over three weeks to get him off the list. Kennedy used a hearing to grill Homeland Security Undersecretary Asa Hutchison: “If they have that kind of difficulty with a member of Congress, how in the world are average Americans, who are getting caught up in this thing … going to be treated fairly and not have their rights abused?” It’s a good question, and one that to this day the government has not answered adequately.12
Even more threatening to our liberty than database screw-ups may be when the database programs work as intended. The government has the means to gather incredible amounts of information on all of us. What it doesn’t collect itself it buys from private vendors. During the Johnson and Ford years the suggestion to “link government databases” was met with public protests. Now it happens without us even knowing. Discussing the TSA’s efforts, a lawyer with the Electronic Privacy Information Center said, “The average person doesn’t understand how much intelligence-driven matching is going on and how this could be accessed for other purposes.”13
The government’s use of data to monitor and control citizens is likely to prove the largest problem ordinary Americans face in this century regarding our personal privacy and security. Yet it is one about which we know and understand little. The difficulty is that, as the EPIC lawyer told The New York Times, all this data collection and data matching is occurring with “no meaningful oversight, transparency, or accountability.” There are statutes that govern privacy and data sharing, but most of these have exceptions for law enforcement, eliminating protection just where we need it most.14
When it comes to the use of private data for law enforcement purposes, the government pretty much makes it up as it goes along. The consequences for anyone caught up in all this can prove staggering. And all of us are having our personal information gathered, often without our permission. What’s needed here is democratic authorization of database systems before they are created, transparency throughout the process, and proper constitutional analysis of those programs that exist. Yet these most basic protections are often missing.
THE RISE OF DATA
Databases have been with us for a long time, but comparing the “then” of scribbled notes stored in a file cabinet to the “now” of Big Data is like confusing the metal tube Copernicus used to see the stars with the Hubble Telescope. Take, for example, the FBI’s legendary fingerprint capabilities. The project was launched in 1924. That database, even bolstered by an early card sorter, largely was searched manually until 1985, when the Bureau’s Automated Fingerprint Identification System (AFIS) allowed what once took thousands of hours to occur in a matter of minutes. Then, in 1999, the database was converted to digital technology, and renamed the Integrated Automatic Fingerprint Identification System (IAFIS). By 2014 the FBI announced that IAFIS itself was about to become obsolete; its NGI—“Next Generation Information—system was “fully operational,” linking more than 18,000 state, local, and federal forces together to share data on a real-time basis. Today, in a matter of minutes the system can compare tens of millions of criminal and civilian records, along with physical characteristics such as mug shots, scars and tattoos, and aliases.15
The database revolution is driven by a trifecta of factors—the proliferation of data, exponential advances in computing speed, and revolutionary miniaturization and growth of storage capability. Today, an Intel microchip is thousands of times faster and tens of thousands of times cheaper than its first chip, manufactured in 1971. Intel’s current CEO, Brian Krzanich, observed that if a 1971 Volkswagen Beetle followed the same trajectory, today it would travel 300,000 miles per hour, and go “two million miles per gallon of gas.” The same is true of data storage. One estimate says 90 percent of the world’s data has been created in the last two years; a similar prediction is that the amount of data stored will double every two years until 2020. You can place the entire printed collection of the Library of Congress on a hard drive that can fit in the palm of your hand.16
Aided by this technology, law enforcement’s databases are vast and growing. The National Crime Information Center currently holds more than 13 million records ranging from gun control background checks to stolen license plates to sex offenders. On a single day in June 2015, it answered over 14.5 million queries. Then there are the terrorism databases. The National Counterterrorism Center’s Terrorism Identities Datasmart Environment (TIDE)—the central database for the intelligence agencies—has ballooned to more than a million names, including some 25,000 American permanent residents and citizens. Every night 1,000 or more new names and bits of information come pouring into the FBI’s Terrorist Screening Center, where analysts begin deciding who ends up on the watch list.17
But it’s not just criminals and terrorist suspects—we are all in the government’s data clutches. The government has our tax data, property ownership histories, car registration records, and employment data. It keeps track of our physical characteristics, travel activities, and can even hold information gleaned from previous government interactions, such as what book somebody was bringing through airport security when they got stopped for special screening. It maintains intelligence and security files on us. Through an “Automated Targeting System,” the Department of Homeland Security (DHS) brings together disparately held data—profiles that are “secret, unreviewable, and maintained by the government for 40 years”—to perform statistical terrorism risk assessments on each person entering or leaving the country.18
One of the most ambitious government data-gathering efforts has been “fusion centers,” which allow state, local, and tribal government law enforcement agencies to join resources and information—ostensibly to prevent terrorist activity. These were created with federal funds after the 9/11 Commission targeted the failure of government officials to share threat information as a primary cause of the 2001 terrorist attacks. In October 2006, Director of National Intelligence John Negroponte explained in remarks at the FBI National Academy that “the federal government can’t be—and should not try to be—everywhere all the time.” Rather, “[o]ur state and local colleagues are our eyes and ears throughout the nation.”19
Although originally about counterterrorism, the mission of most fusion centers quickly morphed to something much broader: “all hazards, all crimes, all threats.” Everything. This is not entirely surprising. As a Sacramento police lieutenant who conducted a Department of Homeland Security study pointed out, there is “insufficient purely ‘terrorist’ activity to support a multi-jurisdictional, multi-governmental level fusion center that exclusively processes terrorist activity.” Analysts, he pointed out, would get bored and their “skills would atrophy.” So the focus of fusion center personnel moved well beyond terrorism. And they share the information they gather widely: with schools, licensing agencies, child-care businesses, and transportation services, even the private sector.20
The amount of citizen information being “fused” is staggering. Delaware State Patrol Captain Bill Harris, who ran the Delaware Information and Analysis Center, explained what it entails: “The fusion process is to take law enforcement information and other information—it could be from the Department of Agriculture, the Department of Transportation, the private sector—and fuse it together to look for anomalies and push information out to our stakeholders in Delaware who have both the right and need to know.” “I don’t want to say it’s unlimited, but the ceiling is very high.” Rhode Island’s Deputy Superintendent of the State Patrol emphasized, “There is never ever enough information … That’s what post-9/11 is about.”21
A critical data entry point into fusion center databases is “suspicious activity reports” or SARs. Cops on patrol, and ordinary citizens, are told to provide a tip if something seems off. These reports get entered into the national “Information Sharing Environment.” Reasonable suspicion of criminal activity is not required for SARs to enter the system; much innocent conduct will land you in the database. The Los Angeles Police Department lists sixty-five activities indicative of “foreign or domestic terrorism” warranting a SAR, including “using binoculars,” “drawing diagrams,” “espousing extremist views,” and “taking pictures or video footage ‘with no apparent esthetic value.’” (Stop and read that last one again.)22
What the government doesn’t collect on you these days, it buys. Private vendors such as Choicepoint and Acxiom have made vast fortunes harvesting and bundling private data for sale to governments. (Choicepoint was purchased in 2008 for $3.6 billion by Reed Elsevier; formerly an educational publisher, the recently renamed RELX Group is making a killing by providing information for risk management.) The companies also run analytics for the government. For example, Raytheon markets an “extreme scale analytics” program called “RIOT”—for Rapid Information Overlay Technology—that enables the government to create a literal map of a person’s life. Pulling photos from social media with embedded location information, RIOT creates charts indicating where you go, and when, while also mapping relationships among people. There’s a promotional video where you can watch Raytheon track one of its own employees (“so now we know where Nick has gone, and we know now what Nick looks like…”)23
The real muscle of law enforcement’s proactive, deterrence-based efforts is data mining—the practice of trying to discern patterns in all this data it has gathered from public and private sources. The number of law enforcement data-mining programs is large and growing. The software program Beware, sold to law enforcement by the company Intrado, integrates “billions of publicly-available commercial records,” to provide “threat scores, headlines and ‘Be Aware’ statements … in a matter of seconds” to police first responders. Given a particular address, Beware can, say, identify the cars registered at the location and provide phone numbers and criminal records of residents; it also searches social media content and online purchases to determine the risk an officer faces.24
What’s all the rage now, at every level of government, is “predictive policing.” Police forces are using software programs with names such as PredPol and HunchLab to try to identify where crime will happen next. “By looking at the whole picture,” says one professor working with the Chicago Police Department, “you can begin to learn what it means for a certain area to be abnormal.” (Study policing data programs for a while and you start to notice this word gets used a lot, this relentless hunt for the “abnormal.”) Similar programs have sprouted up across the country, enabled by enthusiastic corporate partners. IBM developed a program called BluePALMS (Predictive Analytics Lead Modeling Software) that allows the police to enter details of an unsolved crime and “get a list of 20 suspects within one minute.” “This is not science fiction,” boasts Miami’s Lieutenant Arnold Palmer on an IBM video touting the technology.25
All this can get a little spooky, as is evident from the Raytheon video showing how data collection and aggregation allows them to track their employee Nick. The sales pitch moves to its denouement: “Now we want to try to predict where he may be in the future … if you ever did want to try to get ahold of Nick, or maybe get ahold of his laptop, you might want to visit the gym at 6 a.m. on Monday.”26
Imagine you are Nick. You may be, soon enough.
THE HUMAN COSTS OF DATABASES
To be clear, there’s promise in these new technologies. An employee of Intrado, which sells the Beware product, tells of a case where a man took a woman and her daughter hostage. Police wanted to enter into negotiations, but calls to the woman’s home phone went unanswered. Using Beware, they got her cell phone information, which ultimately led to a conversation that diffused the situation.27
But there’s trouble in Big Brother’s paradise, too.
Given the enormous volume of information flowing into government hands, and the low standards for ending up in databases and on watch lists, errors are frequent, with just the sort of consequences Abe Mashal and Senator Kennedy faced. Tens of thousands of travelers are stopped wrongly every year because of incorrect information in the terrorism database.28
Similar problems exist locally, as horror stories about proliferating “gang” files make clear. Three teenage Vietnamese girls were put into a gang database when stopped at a mall in Orange County, California, for dressing in “baggy pants and tight shirts.” One got stopped again; this time—because she was in the database—she had her purse and phonebook searched illegally by a cop who told her, “If you have a problem with this, then don’t come to my city.” Following a shooting in San Diego, a number of young African American men with no criminal records and no complicity whatsoever were jailed because they’d been placed on a confidential law enforcement list that labeled them gang associates. (In San Diego, as elsewhere, you can land in the gang database if police believe you are wearing gang-related clothing or think they have seen you make a gang sign.)29
This lust to gather information on anything that anyone deems “abnormal” or out of place has caused fusion center agents to engage in colossal wrong turns, squandering precious law enforcement resources. Virginia’s fusion center concluded that black colleges were “a radicalization node for almost every type of extremist group.” It advocated surveillance of students. Across the Potomac in Maryland, a nineteen-month investigation into war protesters nabbed two Catholic nuns and a well-known political activist, labeling them “terrorists.” In Missouri, a publication on right-wing militant groups fingered followers of the third-party presidential candidates Bob Barr and Ron Paul as threats, causing a media storm. Nationwide, local, state, and federal authorities have monitored the Black Lives Matter–led protests of police shootings.30
Although these incidents underscore the threat to civil liberties from overzealous data collection, the broader danger to many of us comes from the government, overwhelmed by information, missing what’s critical. In 2007 alone, the Terrorist Screening Center recommended the removal of more than 22,000 names that did not belong in the database. An audit after 2009’s failed Christmas Day airplane terrorist bombing showed that hundreds of people on the no-fly list had been granted visas, which the State Department promptly revoked on the premise that such “individuals could present an immediate threat.” Before he was nabbed, the Times Square bomber—who was on the terror list—got on a plane and sat there until just before takeoff. The failed Christmas bomber, on the other hand, was found unsuitable for the no-fly list despite his father’s warnings to U.S. officials, who had entered the report into TIDE.31
Russ Travers, who ran the government terrorism database, says, “The single biggest worry that I have is long-term quality control.”32
UNREGULATED DATABASES
It’s no wonder quality control is a concern: there is a notable lack of rules regarding the government collecting, storing, analyzing, and sharing our personal data. “[D]ata mining is almost entirely unregulated under current law today,” offers one legal expert. Another says that “in contrast to the approach in many other nations, it is unusual in the United States to find any comprehensive privacy laws.” Even when privacy laws are on the books, those laws almost universally exempt law enforcement.33
In the 2009 case Herring v. United States the Supreme Court pretty much eliminated any Fourth Amendment protection for people who are injured by law enforcement reliance on databases. Bennie Herring was arrested after a computerized database said he had an outstanding arrest warrant. A subsequent search of his person turned up drugs and a gun. But the arrest warrant had been recalled months earlier, and should have been removed from the database; the failure to do so was the result of negligence on the part of police department employees. Some of the justices expressed alarm about the damage such errors could cause, and felt the government should be responsible. “Electronic databases form the nervous system of contemporary criminal justice operations,” wrote Justice Ruth Bader Ginsburg, and “[t]he risk of error stemming from these databases is not slim.” But the majority disagreed; unless “police have been shown to be reckless in maintaining a warrant system, or to have knowingly made false entries to lay the groundwork for future false arrests,” there is not much the Constitution has to say about it.34
On the one hand, you can see where the Supreme Court majority in Herring was coming from. Imagine the impact if the result of any database mistake by law enforcement was that the suspect received a get-out-of-jail-free card. That can’t be right.
On the other hand, with no remedy in law, there’s no incentive for government to do what it takes to get its databases right. As Justice Ginsburg pointed out, private businesses are held liable for employees’ mistakes precisely on the theory that employers will then assure they don’t happen.35
So what is to be done?
Some argue that the only proper restraint on databases is after-the-fact auditing to ensure there was no misbehavior. That’s the view of Stewart Baker, a brilliant lawyer of conservative leanings who served as the General Counsel of the NSA before 9/11 and later was chief of policy at the Department of Homeland Security. Following his tenure in government Baker wrote Skating on Stilts, a take-no-prisoners denunciation of those who oppose security measures in the name of civil liberties. He is bullish on data’s promise, and skeptical about most proposed limitations on data mining.
Baker concedes there’s an “uncomfortable pattern to the use of data by governments.” Government gets data for one reason—be it crime prevention or social security—then, “as time goes on, it becomes attractive to use the data for other, less pressing purposes—collecting child support, perhaps, or enforcing parking tickets.” Still, he says, it is “fighting technology” to try to limit the use of data to the purpose for which it was collected. “It’s like wearing someone else’s dress. Over time, use restrictions end up tight where they should be roomy—and loose where they should be tight.” Similarly, he feels that lawyers, trained in traditional Fourth Amendment law, are just too focused on before-the-fact “predicates” such as the probable cause that the government has to meet to allow it to go a-searching.36
What Baker’s all about—and he is hardly alone in this approach—is management and auditing. “If the lawyer’s solution is to put a predicate between government and the data and the bureaucrat’s solution is to put use restrictions on the data, then … the auditor’s solution” is simply to catch those who use data improperly. “Government access to personal data need not be restricted by speed bumps or walls. Instead it can be protected by rules, so long as the rules are enforced.” (By “rules” he means restrictions on who can use data for what purposes, such as using the database to gather information for personal reasons.)37
The problem is that these sorts of solutions are designed for when things go wrong with databases, not when they go right. Of course we should punish government employees who sneak a peek at what they should not. And we should have mechanisms—much better than we have now—to purge information in the database that is incorrect.38
But after-the-fact auditing is of no help in figuring out how data should be used in the first place. Auditing can do nothing to fix the threat to liberty and security posed by databases holding vast amounts of our personal information that operate all too well.
Take, for example, the urge to collect data to identify communities where trouble might be found. In the 1940s, countless Japanese Americans complied with the United States Census, revealing their personal roots. “[W]ith a growing alliance between Japan and Germany,” writes one author, “these visits could not have been comfortable.” But the Japanese Americans cooperated, perhaps reassured that “by federal law, census data was subject to strict use restrictions” such that personal data could not be revealed. That was the case until 1942, however, when Congress removed those restrictions and the War Department used that information to sweep them into prison camps. There they languished for years, the victims of government taking advantage of data collected for one purpose, and extending it to another. You can see how auditing would not provide much help here. This country has been tripping over itself ever since to apologize.39
Think this can’t happen today? In 2004, the Electronic Privacy Information Center revealed that the Census Bureau provided Homeland Security with a “zip-code breakdown” showing where Arab Americans lived and their “country of origin.” When news of this broke, DHS argued the information was being used to develop Arabic-language signage at airports. Do you believe that?40
Yes, we definitely should figure out a way to address database errors. But we also need to think through why we are creating databases in the first place. Like so much else in law enforcement, when it comes to data we spend too much time mopping up after the fact, rather than thinking through things before we proceed.
NEITHER AUTHORIZED NOR TRANSPARENT: FUSION CENTERS
What’s needed with regard to databases is a return to the basics: up-front democratic authorization and transparency. Too often, when it comes to collecting data, the rule has been to act now, worry about privacy and personal security later. This is exactly backward. Democratic authorization, including clear limitations on the use of data, meaningful privacy policies, and a mechanism for ongoing transparency, should precede establishment of database programs.
Fusion centers are a stunning example of what goes wrong without clear democratic authorization in place. Congress forked out the money for fusion centers, but it did not and could not authorize their activities. “As state and local entities,” a congressional report explained, “the exact missions of individual fusion centers are largely beyond the authority of the federal government.” Unfortunately, most states failed to specify those missions either. Colorado is among the few states that actually defines what a fusion center even is, but Colorado law says little about what a fusion center should do. A 2013 study of fusion centers by the Brennan Center for Justice found “organized chaos.” The federal government had thrown well over a billion dollars at a “loosely coordinated system” of collecting information “with insufficient quality control, accountability, or oversight.” 41
Little surprise, then, that after years of officials lauding fusion centers as a vital tool, a bipartisan Senate investigation essentially labeled fusion centers a joke. When it came to antiterrorism intelligence gathering, the Senate report concluded they were literally worse than worthless. The “intelligence” (the report’s sarcastic quotes) that did flow in was “of uneven quality—oftentimes shoddy, rarely timely.” Not infrequently it simply recapitulated old or publicly accessible news that “more often than not” was “unrelated to terrorism.” “‘[W]hat a bunch of crap is coming through,” said one intelligence chief; another agreed, “You had a lot of data clogging the system of no value.” “[I]n no case,” concluded the report, “did a fusion center make a clear and unique intelligence contribution that helped apprehend a terrorist or disrupt a plot.” Follow that?: “in no case” were fusion centers a help. Worse still, the report flagged three incidents that “raised the possibility that some centers have actually hindered or sidetracked Federal counterterrorism efforts.” 42
On the other hand, the Senate report declared, fusion centers were “endangering citizens’ civil liberties and Privacy Act protections.” DHS was fully aware that information could not be collected or held “solely for the purpose of monitoring activities protected by the U.S. Constitution,” yet such violations were common. A motorcycle gang was investigated for a pamphlet essentially telling “members to obey the law.” One fusion center intelligence report was about a U.S. citizen who gave a motivational talk to a Muslim group on “positive parenting.” Commented a DHS intelligence analyst: “The number of things that scare me about this report are almost too many to write into this [form],” not the least of which was that “the nature of this event is constitutionally-protected activity (public speaking, freedom of assembly, freedom of religion).” 43
This is what happens when government is allowed to violate what should be the cardinal rule of policing: explicit, before-the-fact, democratic authorization. Before government sets out to do something, the something it is going to do should be clear—both to government officials and to the public. The federal government shot fusion centers up with piles of taxpayer dollars for one specific reason—counterterrorism—but the centers themselves decided to take the funds and do something else entirely. Not atypical was Michigan’s fusion center, which “changed its mission,” dropping terrorism entirely and identifying a new goal: to “promote public safety” through a “public-private partnership.” That sort of vague, wide-ranging, self-determined mission is precisely what has led to so many of the problems.44
Equally absent was the other essential half of democratic governance: transparency. Policy analysts who have studied fusion centers say they run the risk of becoming “a ‘one-way mirror,’ in which citizens are subject to ever-greater scrutiny by the authorities, even while the authorities are increasingly protected from scrutiny by the public.” Or, as an official at one fusion center put it bluntly: “If people knew what we were looking at, they’d throw a fit.” An analyst at another center echoed the sentiment, calling it the “wild west,” where operators did what they could “before ‘politics’ catches up and limits options.” 45
CONSTITUTIONAL PROTECTIONS: WHAT’S WRONG WITH OUR DNA TESTING SYSTEM?
Of course, up-front democratic authorization and transparency are just half of what the Constitution demands; when people’s lives, liberty, and property are at stake, the protections of the Fourth Amendment, as well as the Due Process and Equal Protection Clauses, apply. Yet, when it comes to government databases, the analysis of these constitutional safeguards often is extremely muddled—if not neglected entirely.
Government-compelled DNA testing makes this point well. Compared with fusion centers, the country’s DNA database system seems the picture of law enforcement perfection. The accuracy of DNA analysis is highly acclaimed, and congressional and state statutes alike authorize the country’s DNA databases and associated protocols. But the lesson of the DNA testing regime is that even properly authorized databases still must meet the standards of constitutional policing. Our DNA collection system—despite having received the blessing of the Supreme Court—does not. And the reason should be familiar by now: a failure to distinguish suspicion-based from suspicion-less searches.
An Advertisement for DNA Testing
In 2009 a man named Alonzo King was arrested for assault after brandishing a shotgun at a number of people. His DNA was gathered pursuant to a Maryland law mandating collection from anyone arrested for a “crime of violence,” and compared with DNA samples in the cold-case database. That comparison resulted in a match with DNA taken from a violent rape committed some six years previously. King now is serving a prison term of life without parole for that rape.46
King argued that the mandated DNA testing violated his Fourth Amendment rights. The Supreme Court upheld the conviction, touting “DNA testing’s ‘unparalleled ability both to exonerate the wrongly convicted and to identify the guilty.’” It certainly seems hard to argue with a system that puts the Alonzo Kings of this world behind bars for life.47
A Supreme (Court) Lie
While DNA technology is worthy of our admiration, the Court’s opinion in Maryland v. King is not. The decision in King is built on a lie.
The question in King was whether the state can mandate the forcible collection and testing of the DNA of all arrestees. Justice Kennedy’s opinion for the 5–4 King majority said that taking King’s DNA was necessary to “identify” the arrestee. He used the analogy of fingerprints. The state needs to know whom it has taken into custody and—in order to set the terms for pretrial release—whether they have a criminal past.
But, as Justice Scalia pointed out in a dissent dripping with sarcasm at the majority’s disingenuousness, fingerprinting already accomplishes that task. Besides, under existing systems, it took Maryland four months from the time of arrest to the DNA match in King’s case. Identification had nothing to do with why the state of Maryland took and tested King’s DNA.48
As anyone can plainly see, the DNA of arrestees is being checked to solve cold cases. That’s not “identification” as normal people use the word. It is, rather, precisely what Maryland law says: The samples are collected “as part of an official investigation into a crime.” Or, as Maryland’s governor put it, DNA testing “bolsters our efforts to resolve open investigations and bring them to resolution.” 49
What’s at Stake?
You may be thinking, so what? Who cares if the DNA was collected to “identify” King or to solve a crime he committed years earlier? Good for us: they caught the bad guy.
But it matters, because there is much more at stake.
The problem is that there was no probable cause—indeed any cause at all—to believe King had committed the rape, thus justifying collecting his DNA to investigate it. And if there was no probable cause, then the search was flat-out invalid. As we saw in Chapter 7, searches without cause are impermissible unless they are done to all of us. If the government can pick and choose whom to search, without cause, we are all vulnerable.
We’ll examine this point about the lack of probable cause, and what to do about it, more closely in just a moment, but first let’s get a handle on how the use of the DNA database extends well beyond bad guys like King. Consider Elizabeth Haskell, who was arrested in California during a peace demonstration. She was never charged with anything, but officers refused to release her until she submitted a DNA sample. Once in California’s database, it is extraordinarily difficult and costly to get out. So there Elizabeth Haskell sits.50
Elizabeth Haskell is hardly alone: behind our backs, and without any authorization for it at all, the government is building a broad DNA database. DNA is taken from members of our armed services, with the solemn rationale of identifying remains if necessary. But their DNA is kept after they leave the military. Police ask volunteers to give up their DNA in order to crack particularly horrific unsolved crimes. But these volunteers then end up in the database, often without recourse. One person agreed to submit DNA to help solve a murder; it took six years and a lawsuit to get out of the database. Police nab DNA from cigarette butts and Breathalyzer mouthpieces; courts have said that’s fine too. Hundreds of thousands of blood samples from newborns sit in state hands and there is worry these, too, will be added.51
If you are wondering who cares if anyone ends up in the database, the answer—apparently—is that most of us do.
Plenty of respected figures have argued for a universal database, and it is clear from the efforts described above that the government wants one. Law professor Akhil Amar, whom we met in Chapter 3 as author of the argument that the Constitution does not require warrants, wrote on the opinion pages of The New York Times that a universal DNA database would represent “a real improvement in the criminal justice system” given that it would “increase the odds of finding the guilty, freeing the innocent, and vindicating victims.” For that reason, he concluded, “it makes sense to include all citizens in the database.” Even the founder of DNA testing, Alec Jeffreys, himself says that “[i]f we’re all on the database, we’re all in exactly the same boat—the issue of discrimination disappears.”52
But we aren’t going to have a universal database anytime soon, because people just won’t go for it. Popular support is not remotely there for the government collecting and storing all our DNA. And perhaps for good reason.
DNA testing is said to involve the analysis of what is known as “junk DNA.” The appellation “junk” refers to the fact that unlike other parts of the DNA strand, which are highly revealing about a person, the information used in cold-case testing supposedly reveals little except the fact of the match. Scientists have figured out how to gather information from some thirteen sites, or loci, on the DNA strand, to create a profile that—when compared with other profiles—is highly accurate in assuring a match while supposedly telling nothing else about the subject whose DNA is taken.53
The first problem with the “junk” DNA theory is that even the junk is revealing of our personal traits. It can reveal Type I diabetes. It can reveal ethnicity. Indeed, in some places DNA profiles are stored in some databases with racial coding. Jeffreys has suggested that “further troubling links between DNA fingerprints and disease will emerge.”54
But it is worse yet, because if the government has your “profile” built on the supposed “junk,” it likely also has the sample that the profile was constructed from—and that sample holds your entire genome, meaning everything about your genetic makeup. Samples aren’t destroyed after they are used to construct DNA profiles; they are stored away, often into perpetuity. State laws vary and many are silent on the subject of how long.55
We’re assured the original samples won’t be touched—why, then, are they kept?—but that is a blatant falsehood as well. We know this because stored samples already have been accessed for some aspects of “familial” DNA testing. Familial testing works like this: the government runs the profile of an arrestee against cold-case profiles and comes up with a partial match for a murder that occurred years ago. The arrestee isn’t the murderer, but someone in his genetic family is. Knowing this fact, the government can try through further investigation to locate the murderer. In order to facilitate familial matching, California has allowed investigators to go back to original DNA samples to conduct a new test used to confirm male familial ties. Although success stories in familial testing are extremely rare, the government has used this method of familial matching to catch some serious bad guys, such as the notorious BTK (“bind torture kill”) murderer. But this use of stored DNA samples has occurred without any democratic authorization whatsoever. As a result, entirely innocent people get visited by law enforcement and asked to map their families; many of these innocents live for some time under a debilitating “cloud of suspicion.”56
If the government can break into the samples for a purpose not properly authorized by law, and certainly not authorized at the time the samples were taken, they can break in for other reasons down the line. Like when some scientist discovers a way to test DNA for a “predisposition” to crime. That is not as farfetched as it sounds, given that—as the privacy scholar Jeffrey Rosen puts it—“genetic research … reveals increasing ties between genes and predisposition to violence and other antisocial behavior.” Welcome to Minority Report.57
What we do know is that governments have treated cavalierly DNA samples that supposedly were inviolate. When Iceland sold its citizens’ genetic data to deCode Genetics, its founder, Kari Stefansson, said, “We have never claimed that the protection of privacy cannot be broken. The principle element here is trust.” That didn’t cut it with the Icelandic people; now there are stricter controls in place, including jail time for decoding identity. History confirms that governments have a bad record of protecting data. “Perhaps the construction of such a genetic panopticon is wise,” wrote Justice Scalia, dissenting in King. “But I doubt that the proud men who wrote the charter of our liberties would have been so eager to open their mouths for royal inspection.”58
The Costs of Constitutional Dissembling
Which brings us back to the utter failure of constitutional guarantees here. No one disputes the fact that DNA testing is a “search” for Fourth Amendment purposes. Indeed, there are at least two searches: the swab of an arrestee’s cheek to collect the DNA, and the testing of the profile against cold cases.
What there is not in these cases is any sort of cause—suspicion—probable or otherwise, to justify these searches. That’s why Justice Kennedy had to engage in such disingenuousness to reach the result he wanted. He needed a purpose to justify the search, and “identification” was the one he managed to concoct.
As we learned in Chapters 7 and 8, the government can search without any cause at all, but to do so it must meet one of two conditions. Either it must search all of us, in a statute that clearly authorizes the search. Or it must have a statistical basis for searching some subset of us, but not others.
Neither of these is true in the DNA-testing context.
It may possible to develop a constitutionally acceptable statistical basis for some parts of the DNA database. For people who are convicted (not arrested, but convicted) of certain crimes, there may be evidence showing sufficiently high rates of recidivism. One could argue on this basis that collecting their DNA will help deter future crimes once the offenders are released, because they will know DNA might be found to apprehend them. But this doesn’t quite work as a basis for collection from all convicts, as the recidivism for some of them, such as nonviolent drug offenders, is comparatively low. On PBS NewsHour on July 17, 1998, Benjamin Keehn, a public defender in Boston, made the case that “if we are going to take DNA from prisoners because they are at risk [of committing crimes in the future], why shouldn’t we take DNA from teenagers, from homeless people, from Catholic priests, from any subgroup of society that someone is able to make a statistical argument of being at risk?” And there is no statistical case whatsoever for DNA testing of arrestees, who—like Elizabeth Haskell—are entirely innocent of anything, no different from anyone walking the streets.59
Equally obviously, we are not all in the DNA database. It is clear some in government would like a universal database, but there’s no way the voters are supporting one. So government continues to do what it can: passing statutes that pick on easy targets, such as arrestees, and trapping people in the database once it gets hold of their DNA. But that is flat-out unconstitutional.
As is so often the case when the government acts unconstitutionally, it is acting inefficaciously as well. Does it make sense to continue to stuff the so-called offender database so full of names, many of whom are not offenders at all? Looking at California, the well-respected research not-for-profit RAND Corporation concluded that if the goal is resolving unsolved crimes, “it would seem to be a wiser use of California’s resources to devote them to analyzing the backlog of crime-scene evidence.” In other words, do a better job of developing cold-case DNA data. But that is expensive, time-consuming, and exacting. Swabbing the cheeks of people who have done nothing to justify it is easy. So government does the cheap and easy thing, rather than the sensible one, and our privacy sits unguarded in a database that contains the genetic code of a growing number of people.60
DUE PROCESS OF DATABASES
There’s one more part of the Constitution we’ve not really discussed yet: the Due Process Clause. It is basically the guarantee of transparency for our constitutional rights. The Due Process Clause says that no person may be deprived of “life, liberty, or property, without due process of law.” Among other things this means that if the government is doing something bad to you, it has to explain why and give you a chance to respond to its evidence. In legal terms, the government has to give you “notice” and an “opportunity to be heard.”
When it comes to databases, the government rarely accords due process and courts rarely do anything about it. To date, this is a pretty gaping hole in your rights.
In the world of databases, algorithms are king. Algorithms are the formulas that tell the computer what to look for when a database is searched. One needs a way to sort through all that data, and an algorithm provides it. Whether it is predicting who will commit crime, deciding whose taxes to audit, granting welfare benefits, or monitoring Americans’ communications, the algorithms are making the decisions.
The problem is that although the government is now using those algorithms to decide people’s fate based on the pile of information it holds, it likes to keep the algorithms secret. We don’t know what the government is looking for, and we certainly don’t know whether the use of any particular algorithm makes sense. We are kept at bay by database secrecy, particularly when data analysis is being used to our detriment.
This sort of secrecy is completely untenable, because the protocols used to analyze data in databases are prone to mistakes. Even when it comes to DNA testing and the conclusions drawn from it, there are errors. In the summer of 2015, the Texas Forensic Science Commission came to the conclusion that state crime labs were wrongly stating the probability that DNA evidence matched a particular perpetrator—its new protocol required labs to go from saying there was a one-in-a-million chance the DNA had identified the wrong person, to what was more like one in thirty or one in forty. That’s quite the difference. And there are documented cases of people wrongly convicted on DNA evidence, like Josiah Sutton, who was convicted of rape based on DNA, though he was much shorter and lighter than the victim claimed.61
Which brings us back to Abe Mashal and the no-fly list. Abe’s case finally did go to court, and after years of fighting, he now can fly again. It is because a federal judge called the whole thing for what it was: a flagrant violation of Mashal’s most basic constitutional rights. The bad news is that it took five years and huge amounts of time and money to accomplish the constitutionally obvious.62
What countless people on the no-fly list do not get is precisely what due process should afford them: notice and an opportunity to be heard. The government won’t say whether you are even on the list, let alone why. You get to file a “redress form,” but you aren’t told what you are rebutting. This sort of Kafkaesque nightmare should scare all of us, right down to our anklebones.63
Together, the Fourth Amendment and the Due Process Clause could do a lot of work to protect us where databases and algorithms are concerned, far more than courts have recognized. Every step of the government’s use of our data, from collecting it, to holding it, to disseminating it, to analyzing it, is potentially a search. And using that data against us without explaining how the database was used may well violate our due process rights. At present, though, courts pretty much scrutinize only the collection of data, ignoring all the other steps that follow. They rely on the Fourth Amendment’s prohibition against unreasonable searches and seizures, and sideline the Due Process Clause almost entirely when it comes to data aggregation, dissemination, and analysis. That needs to change.
Would complying with both these constitutional clauses be a bother for the government? You bet. But our lives, liberty, and property are on the line. So is common sense. The government fights and fights and fights to keep information secret, only to have it come out that its data programs make no sense, the algorithms do not work, the error rates are staggering. This is not making us safer. Often, it is wasting vast amounts of money on fools’ errands. “Sunlight,” as Justice Brandeis said in a different context, “is the best of disinfectants.” 64
When family tragedy struck Abe Mashal, he sought succor in his faith. When the government ensnared him in its net, he abandoned that faith. His ACLU lawyers told him to do what he wants; the government shouldn’t be allowed to crush the practice of his religion. “That’s true,” Mashal says, resigned, “but you have to live and learn, too.” The question is whether that is the lesson we all should learn from Abe Mashal’s case. That the government can toss us into its databases without sufficient protections, and if our lives are mangled as a result, c’est la vie. Or whether we should get wise, and regulate databases properly.65
* * *
We’re almost at the end of our road. But not quite. We’ve not yet discussed counterterrorism, and what may be the government’s biggest data grab yet, one that captured the nation’s full attention: the use of bulk data collection for national security purposes. That problem, our last, deserves its own chapter.