The New McCarthyism: How The Massachusetts Supreme Judicial Court Got Automated License Plate Readers And The Mosaic Theory All Wrong

By: Dan Noffsinger

INTRODUCTION

In the race to identify rioters at the U.S. Capitol during the attack on January 6, 2021, an often-overlooked technology, nearly invisible in commuters’ daily lives, played a key role—Automated License Plate Readers.^[1] ALPRs^[2] “can do in minutes what it took a cop to do in an entire shift,”^[3] as they are mounted in various locations or on police cars, where they can capture upwards of 2,000 plates per minute.^[4] Even when cars travel in excess of 100 miles per hour, artificial intelligence^[5] extracts the plate number from the photograph with 97–99% accuracy^[6] and records it along with the date, time, and location.^[7] In an instant, ALPR systems can automatically flag vehicles that appear on a “hotlist,” alerting officers to stolen cars or abducted children.^[8] Opponents, however, estimate that as little as 0.2 percent of scans result in a hit and argue that the remainder amounts to indiscriminate, suspicionless mass surveillance that can “violate the rights of entire communities.”^[9] Indeed, in 2018 the NYPD settled the last of three years-long lawsuits concerning its surveillance of Muslim neighborhoods for over one million dollars.^[10] The NYPD admitted years ago to having a picture “of every single car that travels in or out of the city” and the capability to “geo-spatially map each location in the city where a plate reader has spotted the car in the past five years.”^[11] But their surveillance expanded to target Muslims, driving unmarked ALPR-enabled cars past mosques.^[12] Nationwide, ALPR use jumped from 17% of police departments in 2007 to 71% by 2012,^[13] and increasingly cheaper technology in recent years has allowed departments to use even more cameras.^[14] In addition to prolific use by police departments, individual universities are now deploying them on campuses as well.^[15]

Courts have universally agreed that an isolated scan from an ALPR is not an unconstitutional search, primarily because it occurs in public.^[16] But there is a growing theory, dubbed the “mosaic theory” of the Fourth Amendment, that mass data should be analyzed as a whole to determine if an invasive search has taken place.^[17] In April of 2020, the Massachusetts Supreme Judicial Court became the first appellate court to adopt the mosaic theory in relation to ALPRs and rule that prolonged ALPR use would constitute a collective Fourth Amendment search.^[18] This Note demonstrates the legal and practical problems of applying the mosaic theory to ALPRs. Despite sounding promising on paper, the U.S. Supreme Court is unlikely to agree with such an application. This Note proceeds in four parts. Part I provides the legal background of both ALPRs and the mosaic theory leading up to Commonwealth v. McCarthy. Part II dissects the flaws in that opinion with respect to the mosaic theory and ALPRs. Part III further examines the practical limitations of applying and administering the mosaic theory to ALPRs, while Part IV instead proposes viable alternatives to limit the growing reach of ALPRs.

I.ALPRS AND THE FOURTH AMENDMENT

Individuals facing ALPR evidence in court have alleged Fourth Amendment violations, often as an unreasonable basis for a traffic stop, but discrete ALPR use has nearly universally been upheld.^[19] The usual reasoning is that a scan occurs in public so there is no unreasonable search, stemming from the Supreme Court’s holding in New York v. Class,^[20] but some courts have found the conclusion so obvious that they do not include a supporting citation.^[21] The Eleventh Circuit went out of its way to affirm ALPR use in a short, unpublished, per curiam opinion^[22] when it could have ruled on other grounds.^[23] But in the past decade, a new theory has emerged to argue against the aggregated use over time of otherwise-legal technology such as ALPRs, and it has been embraced by some courts.

A. The Mosaic Theory

In the 2012 case United States v. Maynard,^[24] the D.C. Circuit introduced a new approach to Fourth Amendment searches based on aggregation, which Professor Orin Kerr has labeled the “mosaic theory” of the Fourth Amendment and described as “requir[ing] analyzing police actions over time as a collective ‘mosaic’ of surveillance; the mosaic can count as a collective Fourth Amendment search even though the individual steps taken in isolation do not.”^[25] The analogy is to a mosaic in that an individual unicolor tile reveals nothing, whereas the entire mosaic is highly detailed.^[26] The theory’s appeal is that it seeks to protect privacy interests against evolving government surveillance technology when the whole is arguably greater than the sum of its parts.^[27] Maynard applied the theory to a four-week period of GPS surveillance of a car and held that it amounted to an unconstitutional search.^[28] The Supreme Court had direct review of Maynard in United States v. Jones,^[29] and the Court could have addressed the mosaic theory and confronted 21st century technology head-on but instead resolved the case with “18th century tort law” by construing law enforcement’s warrantless placement of the GPS device on the defendant’s car as a trespass.^[30]

While Jones contained no explicit mention of the mosaic theory, multiple scholars including Professor Kerr interpreted the two concurring opinions^[31] as Supreme Court support of the theory.^[32] In the years that followed, several commentators argued that the mosaic theory can and should be applied against ALPRs to find their use unconstitutional.^[33] Until two years ago no appellate court had agreed. But in April 2020, the Massachusetts Supreme Judicial Court handed down Commonwealth v. McCarthy, wherein it embraced the mosaic theory and declared that “the defendant has a constitutionally protected expectation of privacy in the whole of his public movements.”^[34]

B. McCarthy: The Mosaic Theory Applied to ALPRs

Massachusetts’ highest court seized upon a reservation that the Supreme Court expressed in a 1983 surveillance case, United States v. Knotts.^[35] There, the Court upheld the warrantless use of a radio device police used to track a vehicle on a single journey, holding that “[a] person traveling in an automobile on public thoroughfares has no reasonable expectation of privacy in his movements from one place to another”^[36] but noting later that “if such dragnet-type law enforcement practices as respondent envisions should eventually occur, there will be time enough then to determine whether different constitutional principles may be applicable.”^[37] The McCarthy court “indeed [] determined that different constitutional principles govern” than did in Knotts when it concerns a network of ALPRs.^[38]

McCarthy fully and expressly endorsed the mosaic theory, finding it to be “wholly consistent with the statement in Katz that ‘[w]hat a person knowingly exposes to the public . . . is not a subject of Fourth Amendment protection,’ because the whole of one’s movements, even if they are all individually public, are not knowingly exposed in the aggregate.”^[39] The McCarthy court approvingly quoted a passage from the Maynard opinion and concluded that “the whole reveals far more than the sum of the parts.”^[40] After noting that the proper test would be to weigh all of the data collected by the government on the subject—as opposed to strictly what it sought to submit into evidence—and how that was “not possible in the record before [it],” the court nonetheless declared that “[w]ith enough cameras in enough locations, the historic location data from an ALPR system in Massachusetts would invade a reasonable expectation of privacy and would constitute a search for constitutional purposes.”^[41] Specifically, the court found that the state’s “one-year retention period . . . certainly is long enough to warrant constitutional protection.”^[42]

II.AND THE FALLACIES OF APPLYING THE MOSAIC



THEORY TO ALPR S

The McCarthy court’s decision to apply the mosaic theory to ALPRs is flawed in many respects. First, it failed to heed a cautionary statement by the U.S. Supreme Court that it quoted later in its own opinion, that “Fourth Amendment cases must be decided on the facts of each case, not by extravagant generalizations.”^[43] This is so because McCarthy ultimately held that there was no constitutional violation in the case at bar, because on the meager data on the record—four cameras from the ends of two bridges—“the limited use of ALPRs in this case [did] not constitute a search.”^[44] Therefore, the court’s sweeping pronouncement that a one-year period of ALPR surveillance with an unspecified number of cameras would “certainly [be] long enough” to constitute an unconstitutional search was merely dicta.^[45] Further, the McCarthy court erred in arriving at that conclusion in five separate ways: (1) by treating the mosaic theory as if it had been approved by the Supreme Court; (2) by treating ALPR data akin to cell-site location information (CSLI); (3) by finding support for its conclusion in the Supreme Court advanced technology case Kyllo v. United States;^[46] (4) by finding a reasonable expectation of privacy; and (5) by failing to consider the exceptions to the exclusionary rule. Each of these will be discussed in turn.

A. Misplaced Reliance on the Mosaic Theory

First, the McCarthy court failed to flesh out the “different constitutional principles” that it claimed governed the case.^[47] The next section in the opinion, and nearly its entire justification for how “an ALPR system in Massachusetts [could] invade a reasonable expectation of privacy” was both titled and devoted to the mosaic theory,^[48] which is not a constitutional principle, but a theory.^[49] The Supreme Court had the full opportunity to endorse the D.C. Circuit’s mosaic theory, as Jones was a direct review of Maynard,^[50] but as noted by Professor Kerr, the majority “resolved the case without reaching the mosaic theory, and neither concurring opinion gave the issue extensive analysis.”^[51] In fact, the Jones majority did not cite to Maynard a single time outside of establishing the procedural history. Nonetheless, the concerns briefly expressed in the Jones concurrences, despite being dicta, have garnered substantial academic discussion.^[52]

The McCarthy court wrongly implied that the Supreme Court had fully adopted the mosaic theory in the Court’s 2018 surveillance case United States v. Carpenter by using an incomplete line from it as a parenthetical quote, that “individuals have a reasonable expectation of privacy in the whole of their physical movements.”^[53] The first half of the sentence from which that line was excerpted is a critical clarifier: “A majority of this Court has already recognized that . . .”^[54] It is an important distinction that Chief Justice Roberts’ recognition of the previously expressed views of his colleagues from non-binding opinions^[55] did not suddenly place an official stamp of approval on the mosaic theory and turn it into binding precedent.^[56] There are many ways to phrase that sentence, and he easily could have given credence to the mosaic theory by discussing it in a Supreme Court majority opinion for the first time, but instead he confined the reference to a bare factual recitation. Indeed, the closest the mosaic theory came to being named in Carpenter—despite Carpenter’s going “all-in” on the theory in his brief^[57] —is in the title of one of a dozen sources in a single footnote in Justice Thomas’ dissent, sources that he collectively cited as criticism for the Katz test of reasonableness.^[58] The immense difficulties that courts would have in administering a mosaic theory-based test if the theory was adopted (along with other reasons why it should not be) will be discussed in Part III.

B. Flawed Analogy to Cell Phone Location Data

The McCarthy court also failed to distinguish the private nature of the cell-site location information (CSLI) of Carpenter from ALPR data. As noted in Chief Justice Roberts’ opinion, “[c]ell phones continuously scan their environment . . . several times per minute”^[59] and “faithfully follow[] beyond public thoroughfares and into private residences, doctor’s offices, political headquarters, and other potentially revealing locales.”^[60] ALPRs do not follow people into residences or offices; they are purely public, and most, like the ones in McCarthy, are stationary, so they do not provide the precise granular data of GPS monitoring. A scan from an ALPR at a particular cross-street does not let someone conclude that the driver was on the sort of private trip that concerned Justice Sotomayor in Jones, “to the psychiatrist, the plastic surgeon, the abortion clinic, the AIDS treatment center, the strip club, the criminal defense attorney, the by-the-hour motel, the union meeting, the mosque, synagogue or church, the gay bar and on and on.”^[61] ALPRs only periodically record a car’s location in public; its owner’s movements are an inference. In a 2020 ALPR case in the Ninth Circuit, United States v. Yang, the defendant was “unlucky” that the single available read of the car he had rented, of the 5 billion scans in the system, occurred “when he was in possession of the vehicle and was made near his residence,” allowing an officer to locate him.^[62] Judge Bea, in his concurrence, noted that this was a far cry from the 13,000 data points that were collected in Carpenter and fell well short of revealing “particular movements”^[63] or the “familial, political, professional, religious, and sexual associations” that concerned Justice Sotomayor in Jones.^[64] Additionally, a passenger in a vehicle enjoys complete anonymity from ALPRs, as does anyone utilizing any other form of transportation, so a person can conceal their movements from ALPRs if they so desire much easier than they can from facial recognition, for example.^[65]

The Court in Carpenter quoted from additional cases to explicitly contrast cell phones and cars. The Court observed that “unlike the . . . car in Jones, a cell phone—almost a ‘feature of human anatomy’—tracks nearly exactly the movements of its owner”^[66] and also quoted approvingly from Caldwell v. Lewis that “[a] car has little capacity for escaping public scrutiny.”^[67] Despite this distinction and the conclusion that “historical cell-site records present even greater privacy concerns than [Jones’s] GPS monitoring,”^[68] it was still a close question for the Court, as Carpenter was decided 5-4.^[69] Given that ALPRs unquestionably present lesser privacy concerns than GPS monitoring or

CSLI,^[70] the McCarthy court took an unwarranted leap from the Carpenter court’s passing mention of a reasonable expectation of privacy in “the whole of [one’s] physical movements” to the conclusion that McCarthy had such an expectation in the entirety of his public movements alone.^[71]

The critical distinction between CSLI in private areas and ALPR data from public areas survives the omnipresence of ALPRs and the use of AI. This conclusion draws from two unbroken lines of Supreme Court cases, both of which the McCarthy court acknowledged but misapplied—the first regarding observation of automobiles, the second dealing with advancing technology used by law enforcement. The frequently cited Supreme Court support for ALPR use comes from New York v. Class, which the McCarthy court properly quoted in its opinion: “The exterior of a car, of course, is thrust into the public eye, and thus to examine it does not constitute a ‘search.’”^[72] The McCarthy court traced Class to a “what is knowingly exposed” principle first espoused in Katz v. United States^[73] and found further support in Knotts as well as a 2002 Massachusetts appellate case.^[74] The McCarthy court then veered away from this foundation to embrace the mosaic theory, quoting at length from Maynard,^[75] even though the Supreme Court did not explicitly endorse the Maynard opinion or the mosaic theory, despite having the full opportunity to do so.^[76] The McCarthy court provided only the above partial quote from Carpenter^[77] as support for its assertion that “the United States Supreme Court [has] recognized a privacy interest in the whole of one’s public movements,”^[78] but if it had also analyzed an additional Supreme Court case that followed Knotts the next year, that should have led the McCarthy court away from its conclusion.

In 1984, the Court narrowed the holding of Knotts in United States v. Karo.^[79] After Knotts upheld the warrantless use of a radio beeper to track a car on a public highway, the government sought to admit evidence based on another beeper, similarly placed in a container of chemicals, that was carried into a house, wherein the government was able to verify that it remained hours later.^[80] The Court struck down the new use, distinguishing Knotts because the new information came from “a private residence, a location not open to visual surveillance.”^[81] The Court reasoned that the information in question in Knotts was “voluntarily conveyed to anyone who wanted to look,”^[82] whereas in Karo, “the monitoring indicated that the beeper was inside the house, a fact that could not have been visually verified.”^[83]

As noted above, McCarthy seized upon the reservation in Knotts,^[84] but it did so without ever discussing Karo, which would have revealed the distinction between public and private. While the oft-quoted line from Katz is that “the Fourth Amendment protects people, not places,”^[85] Justice Harlan, in the lead-in of his famous concurrence, noted that determining “what protection it affords to those people [generally] requires reference to a ‘place.’”^[86] This has led to the critical phrase “constitutionally protected area” in Fourth Amendment jurisprudence.^[87] In United States v. Graham, the District Court of Maryland, ruling on CSLI before Carpenter but after Jones, after discussing Karo found it of critical importance that in Graham “[d]efendants ha[d] not argued that the historical cell site records revealed their movements in protected areas such as their homes.”^[88] Carpenter itself, in Chief Justice Roberts’ own words, was “about a detailed chronicle of a person’s physical presence compiled every day, every moment.”^[89] Before arriving at this conclusion, Chief Justice Roberts quoted from the concerns in Knotts about “twenty-four hour surveillance” earlier in the opinion.^[90] ALPRs do not chronicle every moment, twenty-four hours per day, and they do not intrude on any constitutionally protected areas. The two aforementioned ALPR cases from 2020 evidence this—over multiple months, McCarthy was scanned only on a single bridge,^[91] while Yang’s rental car registered just once in over a week.^[92]

Karo dovetails with the second line of cases that McCarthy misapplied, concerning the use of advancing technology by law enforcement.^[93] The McCarthy court set the tone early on for its conclusions by quoting from its own recent case that “both this court and the United States Supreme Court have been careful to guard against the ‘power of technology to shrink the realm of guaranteed privacy’ by emphasizing that privacy rights ‘cannot be left at the mercy of advancing technology but rather must be preserved and protected [from] new technologies’”^[94] of the government. As Supreme Court support for this assertion, the McCarthy opinion cited to the 2001 case of Kyllo v. United States, without providing any detail.^[95]

C. Inappropriate Reliance on Kyllo

Kyllo is instructive and did anticipate advancing technology, but when read fairly, it instead provides support for ALPR use and cuts against the McCarthy court’s conclusions. There, the Court struck down federal agents’ use of thermal imaging performed from a public street but directed into the defendant’s house, which the agents used to show that he was growing marijuana inside.^[96] The Court held that “[w]here, as here, the Government uses a device that is not in general public use, to explore details of the home that would previously have been unknowable without physical intrusion, the surveillance is a ‘search’ and is presumptively unreasonable without a warrant.”^[97] The Government sought to rely on Dow Chemical Co. v. United States, a case involving aerial visual surveillance that, the Kyllo court acknowledged, held “that visual observation is no ‘search’ at all.”^[98] The Kyllo court noted that visual surveillance was largely unquestioned dating back to English common law and recognized Dow Chemical when it stated that “technology enabling human flight has exposed to public view (and hence, we have said, to official observation) uncovered portions of the house,”^[99] but it held that thermal imaging was a bridge too far in that it was much more than a force-multiplier because thermal information “would previously have been unknowable without physical intrusion.”^[100]

Parsing the holding of Kyllo, its conclusion, shows how it contradicts McCarthy’s reasoning regarding ALPRs at each of three turns: (1) “not in general public use;” (2) “details of the home;” and (3) “previously unknowable without physical intrusion.”^[101] First, ALPR technology is readily available to the public, starting in 2015, free software could turn any internet-connected camera into an ALPR.^[102] For example, the technology is becoming popular with the likes of homeowners associations^[103] and property managers.^[104] The “general public use” test has been criticized as a “loophole” that is difficult to administer.^[105] However, it necessarily looked to the future,^[106] and further, “[t]he touchstone of the Fourth Amendment is reasonableness,” harkening back to the expectation of privacy test of Katz.^[107] After all, if it was commonplace for people of the 1960’s to carry around parabolic microphones, it would not have been reasonable for Mr. Katz to expect privacy in his famous phonebooth.^[108]

Secondly, the Kyllo opinion afforded strong protection to the residence, noting in its very first sentence that the surveillance was gathered from a private home.^[109] The Court stated that “‘[a]t the very core’ of the Fourth Amendment ‘stands the right of a man to retreat into his own home and there be free from unreasonable governmental intrusion’”^[110] and declared that “[w]ith few exceptions, the question whether a warrantless search of a home is reasonable and hence constitutional must be answered no.”^[111] There are no such heightened concerns in the case of ALPRs stationed in public, making their use akin to the enhanced aerial photography of Dow Chemical, where the Court found, as noted by the Kyllo court, “it important that this is not an area immediately adjacent to a private home, where privacy expectations are most heightened.”^[112] This remains true even if the car was parked in a driveway or an open garage.^[113]

Lastly, the Kyllo court placed a significant distinction on the fact that the thermal imaging “would previously have been unknowable without physical intrusion.”^[114] The Court reasoned that finding a search on those grounds “assures preservation of that degree of privacy against government that existed when the Fourth Amendment was adopted.”^[115] The entire purpose of a license plate is to make the vehicle identifiable to law enforcement, and thus as noted by the Class court decades ago, “it is unreasonable to have an expectation of privacy in an object required by law to be located in a place ordinarily in plain view from the exterior of the automobile.”^[116] The McCarthy court, however, incorrectly placed ALPR use in the “previously unknowable” category with CSLI, stating that “[l]ike CSLI data, ALPRs allow the police to reconstruct people’s past movements . . . thus granting police access to ‘a category of information otherwise unknowable.’”^[117] The McCarthy court also emphasized, from its own past case, that “the government, without securing a warrant, may use electronic devices to monitor an individual’s movements in public to the extent that the same result could be achieved through visual surveillance.”^[118] Curiously, McCarthy used this precedent to argue against ALPR use, when it should have pointed to upholding the use.

Here, the McCarthy court failed to recognize the force-multiplier nature of an ALPR network.^[119] It admitted that “an officer may read or write down a publicly displayed license plate number. In this way, a single license plate reader is similar to traditional [allowable] surveillance techniques,” but then listed four factors that it considered distinguishing: retention, recording nearly every vehicle, its ongoing nature, and the inclusion of a location.^[120] The first and final factors can be easily disposed of, as a single police officer can readily write down a car’s location and retain it.^[121] The middle two factors also distinguish ALPRs from CSLI, because a finite number of officers working in shifts could duplicate the work of a finite number of ALPRs in public spaces, whereas no number of officers could produce the comprehensive record of CSLI.^[122] Therefore, an ALPR network should be allowable because it “merely augment[s] officers’ physical abilities” as in Knotts,^[123] and at all times, it constitutes solely “visual observation [which] is no ‘search’ at all.”^[124] For all these reasons, Kyllo fully supports the use of ALPRs.

D. The Reasonable Expectation of Privacy

While concurring opinions from Jones cited to Kyllo in expressing their concerns with aggregated data,^[125] those concerns should be allayed when it comes to ALPRs. Such searches are always made in public and are consistent with a reasonable expectation of privacy, which remains the heart of Fourth Amendment search jurisprudence.^[126] As described in 2018 in Carpenter, ever since Katz v. United States,^[127] “[w]hen an individual ‘seeks to preserve something as private,’ and his expectation of privacy is ‘one that society is prepared to recognize as reasonable,’ we have held that official intrusion into that private sphere generally qualifies as a search.”^[128] In Jones, Justices Sotomayor and Alito, speaking for five justices between them, both indicated that they would have resolved the case with the reasonable expectation of privacy test.^[129] The mosaic theory, not explicitly mentioned in any portion of Jones, would be a means to that end. Even those that argue for the adoption of the mosaic theory to strike down ALPR use must concede that ALPRs are significantly less invasive than GPS monitoring,^[130] which involves “constant, uninterrupted monitoring”^[131] even into private areas of the sort that concerned Justice Sotomayor in Jones.^[132] Cameras are also more expected—drivers have long been aware that there are cameras on traffic lights (“red-light cameras”) and toll booths, for example, to accompany police officers.^[133] Given that most medium-sized American police departments have ALPRs,^[134] most regular drivers have experienced ALPR surveillance and should therefore expect it.^[135] In McCarthy, “a testifying expert alluded to cameras ‘all over the State,’”^[136] and amici submitted that there were over 150 cameras five years before.^[137] Massachusetts drivers, like McCarthy, had plenty of notice of the existence of the cameras in the state certainly by the time the Boston Globe, among others, reported on the Boston Police Department’s halting of its ALPR system in 2013 after public records requests revealed that the department was not following its own policies and not properly protecting the data.^[138]

The Court has found a reasonable expectation of privacy to be lacking based on much more infrequent occurrences. In California v. Ciraolo, officers responding to an anonymous tip “secured a private plane and flew over respondent’s house at an altitude of 1,000 feet,” and from that vantage point they could identify marijuana growing in the defendant’s backyard,^[139] at least with the aid of a standard 35mm camera.^[140] The defendant clearly manifested a subjective expectation of privacy by erecting a 10-foot inner fence,^[141] but the Court reasoned that “[a]ny member of the public flying in this airspace who glanced down could have seen everything that [those] officers observed” and therefore “readily conclude[d] that respondent’s expectation that his garden was protected from such observation is unreasonable.”^[142] In a point applicable to ALPRs, the Court’s majority opinion countered a contention by the dissent by stating that “Justice Harlan’s observations about future electronic developments and the potential for electronic interference with private communications were plainly not aimed at simple visual observations from a public place.”^[143]

Three years later, the Court considered a similar case, this time involving a helicopter, in Florida v. Riley.^[144] Officers again observed marijuana in a backyard after responding to an anonymous tip, this time from a helicopter at a height of 400 feet, lower than what is allowed by fixed-wing aircraft like those at issue in Ciraolo.^[145] But since helicopters are allowed at that height, the Court reasoned that “[a]ny member of the public could legally have been flying over Riley’s property,” and it set the threshold frequency of such an occurrence quite low, observing that “there is no indication that such flights are unheard of” in the region.^[146] ALPR systems are far from unheard of, as most medium-sized police departments in the country utilize them,^[147] and many agencies that employ such a system give full public notice of its use, often accompanied by newspaper articles.^[148]

Even if a court finds a search in a particular case, there is an additional step, as the Fourth Amendment protects only against “unreasonable searches.”^[149] The argument for a reasonable search was not considered in Jones because it was not raised below and thus forfeited,^[150] but the analysis in a subsequent car GPS tracking case from the Court of Appeals of New York is revealing. In Cunningham v. New York State Department of Labor, a New York state employer suspected an employee of falsifying time sheets, so it attached a GPS to his car.^[151] Relying on Jones and Weaver,^[152] New York’s highest court found that this constituted a search, but it then considered whether the search was reasonable given the suspicions that the employer had.^[153] The court held that the twenty-four-hour monitoring, including Cunningham’s vacation, was unreasonable, but it strongly suggested that if the scope of the monitoring had been limited to business hours, the search would have been reasonable.^[154] Applying this rationale to ALPRs, monitoring occurs only when people are actively out on public roads, and then only intermittently,^[155] so any search that could be found by a court should still be reasonable and thus allowable.

E. Failure to Consider Exceptions to the Exclusionary Rule

Lastly, the McCarthy court was too quick to apply the mosaic theory given the case’s procedural posture. ALPR validity was at issue due to a motion to suppress the ALPR evidence.^[156] ALPR use is commonly challenged directly at that stage, or similarly as an improper basis for a warrant.^[157] In either scenario, the exclusionary rule is a doctrine that courts can apply to block the government from entering evidence that was obtained in violation of the Constitution.^[158] However, in Stone v. Powell, the Supreme Court stated that the exclusionary rule “is not a personal constitutional right,” as its “primary justification . . . is the deterrence of police conduct that violates Fourth Amendment rights.”^[159] The Court reasoned that “this concern has limited force as a justification for the exclusion of highly probative evidence.”^[160] The ability to place a subject at the scene of a crime via an ALPR is, of course, highly probative evidence. Furthermore, there is Supreme Court precedent that on such a motion to suppress on grounds of an unconstitutional privacy intrusion, the burden lies with the defendant.^[161] McCarthy’s counsel admitted that he “possesse[d] the burden to show that he” met both the subjective and objective elements of the Katz test.^[162] Therefore, any doubt about allowing such evidence should have been resolved in favor of the prosecution.^[163] Two exceptions to the exclusionary rule further demonstrate how McCarthy’s rationale is untenable.

First, the good faith exception^[164] should have applied to the officers in McCarthy had the court ruled that there was a search. That exception serves to admit evidence gathered by officers who were operating on the objectively reasonable and good faith belief that they had the proper legal authority.^[165] Given that McCarthy was the first appellate state court to rule that ALPR use could be unconstitutional, naturally the officers were operating on the understanding that the use of the cameras in question, installed two years prior to the events of the case^[166] and controlled by state regulations,^[167] was constitutional. Indeed, for background, the McCarthy court cited a prior Massachusetts appellate case for its holding “that the defendant had no reasonable expectation of privacy that would prevent an officer from examining his license plate.”^[168] It follows that the good faith exception should have applied if there was a search, as the officers’ actions were objectively reasonable in light of existing precedent. However, the McCarthy court did not address this argument, after the district attorney did not raise it in his brief.^[169]

In United States v. Graham, the government did raise the good faith exception to backstop its argument concerning the admission of CSLI data.^[170] The 2012 case came before Carpenter but after Jones, and while the District Court of Maryland found no constitutional violation in the collection of the CSLI, it nonetheless analyzed the exception under the heading “Suppression Would Not Be The Remedy.”^[171] There, the officers acted in reliance on the Stored Communications Act and Magistrate Judge orders that issued under it, and the court found that this was objectively reasonable.^[172] The court concluded, perhaps anticipating Carpenter, that “[e]ven if the government’s acquisition of historical cell site location records in this case had been in violation of the Defendants’ Fourth Amendment rights, it obtained those records in good faith reliance,” and it denied the motion to suppress.^[173] Similarly, Mr. Carpenter’s Supreme Court victory was short-lived, as his conviction was affirmed in 2019 on remand to the Sixth Circuit, which applied the good faith exception.^[174] The officers in McCarthy should have been afforded similar protection if that was necessary.

The mechanics of the good faith exception show how McCarthy has only muddied the water for future cases in this area. In Davis v. United States, the Supreme Court held that “searches conducted in objectively reasonable reliance on binding appellate precedent are not subject to the exclusionary rule.”^[175] The McCarthy court’s pronouncement that constitutional protection would “certainly” be warranted prior to hitting the state’s one-year retention period^[176] was dicta, meaning that it did not establish new binding appellate precedent that would serve as official notice to officers and limit the use of the good-faith exception. Far from developing a bright-light rule, McCarthy failed to suggest any kind of test for when, in between the six weeks at bar and the state’s one-year retention period, the constitutional line would be crossed.^[177] Thus, it did not provide guidance for law enforcement to regulate its action, missing the mark of Stone and parking itself on a very slippery slope. Since McCarthy did not disturb the prior holding concerning isolated license plate readings,^[178] instead applying the mosaic theory to aggregated data, Massachusetts law enforcement should still be able to use the lack of a clearly established precedent^[179] to rely on the good-faith exception in the next case with similar facts. Further, it would arguably apply each time that a court wanted to lower the threshold.

A second exception to the exclusionary rule demonstrates another flaw in the mosaic theory as applied to ALPRs. Under the Independent Source Doctrine, evidence obtained from an unlawful search may later be admissible if it is obtained through a separate, constitutional search.^[180] This doctrine was endorsed by the Supreme Court in Nix v. Williams under the rationale that the police should be put “in the same, not a worse, position that they would have been in if no police error or misconduct had occurred.”^[181] Thus, if the local police in McCarthy had asked a homeowners association or property manager near the bridge if they could help with crime detection by turning over their ALPR data,^[182] any information they obtained in this manner would be freely admissible in court.^[183] In fact, some such private entities already voluntarily share their data with law enforcement.^[184] Furthermore, if instead the Massachusetts State Police had monitored McCarthy for a full year state-wide, assembling an unconstitutionally large dataset in the eyes of the McCarthy court, they apparently could have turned to the local Barnstable police who actually accumulated the allowable (and relevant) data over two-plus months.^[185] The McCarthy opinion fails to grapple with these practical intricacies.^[186]

III. PRACTICAL LIMITATIONS OF THE MOSAIC THEORY

Only the mosaic theory could defeat ALPR use when each individual piece of information is legitimate on its own.^[187] Even numbers added together can only produce even numbers, no matter how many, but the McCarthy court essentially reasoned that at some point, the sum becomes odd.^[188] In addition to the specific flaws in the McCarthy opinion, courts and scholars have identified additional practical limitations of applying such an aggregation theory generally. Such criticisms also counsel against application of the mosaic theory to ALPRs. Several courts have considered the mosaic theory in Fourth Amendment search cases and found “that approach to be problematic.”^[189] The criticisms from courts and academics can largely be divided into three categories: (1) creating retroactive unconstitutionality; (2) forcing arbitrary line-drawing by courts; and (3) making unworkable guidelines for law enforcement.

A. Retroactive Unconstitutionality

The Graham court noted that the Supreme Court’s majority opinion in Jones did not approve of the D.C. Circuit’s mosaic theory, and Graham instead agreed with Professor Kerr’s objection to “the bizarre consequence of creating retroactive unconstitutionality.”^[190] Retroactive unconstitutionality is illustrated by an extension of McCarthy, which held that four cameras over two months did not create an unconstitutional search but stated that a one-year period would.^[191] Therefore, if the police had monitored McCarthy for an additional ten months after obtaining reasonable suspicion (perhaps in pursuit of a proverbial bigger fish) but sought to enter into evidence only the initial period, the McCarthy court would apparently have ruled that the entire “mosaic” was an unconstitutional search,^[192] including the relevant period that they actually held was allowable. As summed up by the Graham court, “the law as it now stands simply does not contemplate a situation whereby traditional surveillance becomes a Fourth Amendment ‘search’ only after some specified period of time—discrete acts of law enforcement are either constitutional or they are not.”^[193]

B. Arbitrary Line-Drawing

In applying the mosaic theory, even without considering retroactive unconstitutionality, courts would still need to determine, based on the facts of each case, whether the imaginary line of unconstitutionality was crossed. Opinions such as Maynard and McCarthy confidently state that an intrusion did occur (or could, respectively), but neither lays out any rules or methods for evaluating closer cases. This is a common thread, as the mosaic theory is much easier to appeal to than it is to flesh out and apply. As Professor Kerr, who has followed and debated the mosaic theory since its first connection to the Fourth Amendment,^[194] wrote during the Carpenter proceedings, “one of the fascinating aspects of the mosaic theory is that its proponents generally go to great lengths to avoid explaining how they would implement it.”^[195] The reason for this, he feels, is that to fully implement it, it would be necessary “to make a few dozen essentially arbitrary line-drawing calls.”^[196]

Proponents of the mosaic theory may point out that there is no bright-line rule regarding the allowable duration of a Terry stop, for example, but there are far fewer factors at play there.^[197] To evaluate an ALPR dataset on a particular defendant would require considering: the available history of the database, the number of cameras that recorded the suspect, the number of cameras that did not record the suspect, the number of distinct locations, the number of trips in a limited amount of time such that one could reasonably infer that the vehicle went directly from point A to point B, the times of day, etc. The Supreme Court has emphasized how it “repeatedly has acknowledged the difficulties created for courts, police, and citizens by an ad hoc, case-by-case definition of Fourth Amendment standards to be applied in differing factual circumstances.”^[198] A concern the Court has expressed is that such an approach “also creates a danger that constitutional rights will be arbitrarily and inequitably enforced.”^[199]

Furthermore, since ALPRs do not directly track people, only their vehicles, any bright-line rule is subject to manipulation. A criminal who suspects that an ALPR camera placed him near the scene of a crime could intentionally keep driving his car past various ALPRs to try to accumulate the threshold number of scans before evidence was finalized for trial. And if a court was to articulate exactly where it was drawing the lines, it would “quickly become[] clear that you’re really drafting a statute.”^[200]

C. Failing to Balance the Needs of Law Enforcement

The fact that the McCarthy court was not able to illustrate when or where that line might be crossed, even ex post with the benefit of hindsight, illustrates how difficult it would be for law enforcement to determine ex ante when a warrant would be required.^[201] In the 2014 case of Tracey v. State, Florida’s Supreme Court squarely considered the mosaic theory in the context of CSLI and found it to be “not a workable analysis” for similar reasons.^[202] The Tracey court found that applying it would “require[] case-by-case, after-the-fact, ad hoc determinations whether the length of the monitoring crossed the threshold . . . The [U.S.] Supreme Court has warned against such an ad hoc analysis.”^[203] The court, citing U.S. Supreme Court precedent, reasoned that privacy concerns must be balanced with the needs of law enforcement to create “workable rules.”^[204] The Tracey court also acknowledged the concurrences in Jones but recognized that they were dicta and, therefore, “the concerns and questions raised by the concurring Justices were not answered.”^[205] Tracey was decided years before Carpenter, but both holdings agreed that acquiring CSLI was a search, without resorting to the mosaic theory.^[206]

Some believe that Carpenter’s ruling against accumulated data creates trouble for ALPR^[207] by moving away from Graham’s discrete acts and thus indeed changing “the law as it now stands . . . whereby traditional surveillance becomes a Fourth Amendment ‘search’ [] after some specified period of time.”^[208] But there was a discrete act in Carpenter, when the government requested the CSLI from the defendant’s phone provider.^[209] Chief Justice Roberts’ Carpenter ruling contains a clear ex ante instruction to law enforcement regarding CSLI: “the Government’s obligation is a familiar one—get a warrant.”^[210]

Any ex post ALPR analysis is further compounded by the fact that various agencies swap their ALPR data with each other, sometimes without even knowing that they are doing so.^[211] This takes the above scenario to a new dimension, in that access to another government agency’s database—the neighboring towns, for example—could, in the eyes of the McCarthy court, push the collection over the imaginary line well after the fact, perhaps even while a case was on appeal. This would be a nonsensical result, as there would be no new material evidence that should have any bearing on the case.

In all, Professor Kerr^[212] was prescient when he first considered the mosaic theory back in 2010 and stated, “I don’t see what principles there are that could keep it from becoming an extraordinary mess.”^[213]

IV. NON-JUDICIAL ALTERNATIVES TO BETTER BALANCE ALPR USAGE

The mosaic theory should not be adopted by the courts to strike down ALPR usage for all the reasons listed above. Doing so would move away from decades of Fourth Amendment jurisprudence and, by threatening a network of thousands of ALPRs across the nation, amount to legislating from the bench. If the documented crime-solving successes of ALPR^[214] do not outweigh the privacy concerns in the eyes of society, then that is something that legislators can readily address. Individual agencies can also take steps to regulate themselves to quell public concern. On the other hand, parking regulations introduced to effectuate ALPR use have introduced new concerns about safety. This part addresses each non-judicial alternative in turn.

A. Data Retention Limits by Legislation

If this method of surveillance is something that society regards as not worthwhile, the practical solution is legislative. In one of the Jones concurrences, Justice Alito agreed when he wrote that “[i]n circumstances involving dramatic technological change, the best solution to privacy concerns may be legislative.”^[215] The late Justice Scalia, the author of the Jones majority opinion who often favored state’s rights and a legislative solution,^[216] would likely agree. One in three states has taken action in this area, as at least sixteen have statutes directly addressing ALPRs.^[217] However, several of those statutes merely mandate certification, create public records exceptions, or require published policies from the agencies. Only ten states expressly set a data retention limit by statute,^[218] which is what would truly limit law enforcement in favor of privacy protection. States could also give a voice directly to the people—underreported amid the turmoil involving the 2020 presidential election were pro-privacy ballot measures (unrelated to ALPR) passed in California and Michigan.^[219]

The restrictions posed by states vary tremendously. Georgia’s retention limit is thirty months,^[220] which would still be problematic for the McCarthy court, while New Hampshire requires deletion within three minutes in the absence of a hit.^[221] Meanwhile, Arkansas was one of the first states to entirely ban private use.^[222] While the level of national awareness and attention has not been as high as that of sports gambling,^[223] for example, the trend is towards the majority of the states considering legislating and regulating ALPR use, as multiple states have proposed new bills since the start of 2019.^[224] Interestingly, a Massachusetts bill, which would have required law enforcement agencies to delete all ALPR data within forty-eight hours of capture, passed a joint committee in February 2020, two months before the McCarthy opinion, but then stalled.^[225]

B. Tiered Use of ALPR Data

There is an existing federal statute, at issue in Graham, that could serve as a template for ALPR—the Stored Communications Act.^[226] At the outset, it should be noted that while Carpenter clearly affected the Act’s application, by invalidating a warrant that issued from it, the Court’s “opinion did not invalidate [the court order section of the Act] whole cloth.”^[227] There is a framework in the Stored Communications Act that could serve as the basis for a national ALPR statute. Notably, Section 2703(d) of the Act calls for “specific and articulable facts showing that there are reasonable grounds to believe that the contents of . . . the records or other information sought, are relevant and material to an ongoing criminal investigation” before a warrant will issue.^[228] An analogous ALPR statute could still allow real-time “hotlist” monitoring^[229] but require law enforcement to show grounds before a neutral magistrate before accessing historic information that might reveal a pattern of movements of the sort that concern proponents of the mosaic theory. Also of note, language similar to the current limitation in the Act that “a court order shall not issue if prohibited by the law of such State”^[230] could serve to allow each state to set ALPR data retention limits of its own choosing.

C. Individual Agency Adjustments and Oversight

If state legislators cannot reach a consensus to balance privacy with ALPR use, individual municipalities and agencies are free to set their own restrictions, and many have. For example, Minnesota’s overall state limit for retaining license plate data is sixty days, but the Minnesota State Patrol’s is forty-eight hours.^[231] Compare this to the NYPD’s five years.^[232] In response to criticism on Fourth Amendment grounds, the NYPD Deputy Commissioner—apparently briefed on jurisprudence in this area—stated that he did not think that their system “violates anyone’s expectation of privacy.”^[233]

Lobbying efforts to resist ALPR regulation have reached as high as the federal level,^[234] but despite this, some oversight is finally coming to the NYPD.^[235] First introduced in 2017 and met with “fierce opposition” from the NYPD,^[236] the Public Oversight of Surveillance Technology (POST) Act finally passed the New York City Council in June of 2020.^[237] The Act required the NYPD to publish policies by January of 2021 that describe all of their surveillance technologies (including ALPR), their oversight mechanisms, and the procedures to prevent abuse.^[238] Major cities such as San Francisco and Seattle have already passed more stringent laws than the POST Act, and some jurisdictions require outside approval before acquiring new surveillance technology.^[239]

Departments that do employ ALPRs should take care that they do so evenly. A 2020 Buffalo TV news investigation discovered that other than a few on the international border, the remainder of the city’s readers “are laser focused on Buffalo’s east side,” with ten percent of the over 40 million reads from the prior year coming from just two streets.^[240] A Buffalo police captain defended the deployment, stating that violent crime is disproportionate on the east side.^[241] Disproportionate policing, however, is likely to only exacerbate the discrepancy because of what is known as collider bias—“if there’s bias in who the police choose to interact with—if it’s not a random sample—that can change the relationships you see in the data.”^[242] Predictive policing has been criticized for amounting to racial profiling,^[243] and here, recent evidence shows that ALPR use leads to disproportionate attention on poor and minority communities.^[244] The Buffalo Common Council President said that he is looking into it, as he agreed that ALPR cameras “should be equally distributed across the city.”^[245]

There are additional steps that agencies can take to ensure the privacy of its citizens and safeguard against misuse.^[246] For example, if officers insist on being able to look back multiple months, the agency could set a soft deadline of one month, at which point the older data would be migrated to an offline server that could only be accessed from the main office.^[247] Further, that server could be configured to require a tracked login that registers an articulated statement of purpose for each query, at the level of reasonable suspicion.^[248] Also, once that soft deadline is hit, inter-agency sharing should cease as well, in the form of removal from any other database. Finally, at the officer level, all ALPR users on the ground should visually verify the plate for a match to avoid mis-reads.^[249]

D. Ripple Effects for Vehicle Safety

In states around the country that do not require front license plates, enhancing the efficacy of mobile ALPRs has led to new regulations that prohibit back-in parking.^[250] One Florida university that enacted this policy in 2019 in conjunction with its introduction of ALPR^[251] was met with resistance in the form of a petition that has been signed by over 1,500 people, even though the petition did not mention ALPRs.^[252] Rather, it raised safety concerns in slowing exits from structures at nighttime, and it referenced a study that estimated that hundreds of deaths annually and thousands of injuries result from nose-in parking.^[253] Since such parking regulations are only in place to reveal license plates, the twenty states that do not have a front license plate requirement^[254] would be well-advised to allow, if not mandate, a duplicate front plate to give drivers an option to avoid nose-in parking.

CONCLUSION

ALPR use is becoming commonplace in our society, as new locales continue to adopt them on a regular basis.^[255] The mosaic theory is an appealing theory to protect citizens against evolving and encroaching government surveillance, but ALPRs are not the appropriate technology against which to wield it. Scans occur only in public once people have voluntarily ventured out in their vehicles, where there is no expectation of privacy, no matter the accumulation over time. To attempt to draw a line would be as difficult as predetermining how many snowflakes it takes to become a snowball, and any kind of subjective test would unduly hinder law enforcement’s efforts to use this valuable technology. Fortunately, there are viable alternatives, available through statutory reform and regulatory precautions, that can address privacy concerns while allowing ALPR use.

Footnotes

1. Drew Harwell and Craig Timberg, How America’s Surveillance Networks Helped the FBI Catch the Capitol Mob, WASH. POST (Apr. 2, 2021, 9:00 AM), https://www.washington post.com/technology/2021/04/02/capitol-siege-arrests-technology-fbi-privacy/ [https://perma.cc/ R5ZZ-62PP].

2. Different sources use ALPR and LPR, with and without an apostrophe in the plural form, nearly interchangeably.

3. Tod Newcombe, States Start Restricting Police License Plate Readers, GOVERNING (Aug. 12, 2015, 5:00 PM), https://www.governing.com/columns/tech-talk/gov-automated-license -plate-readers-police.html [https://perma.cc/VP2G-KU9E].

4. Justin Rohrlich, In Just Two Years, 9,000 of These Cameras Were Installed to Spy on Your Car, QUARTZ (Feb. 5, 2019), https://qz.com/1540488/in-just-two-years-9000-of-these- cameras-were-installed-to-spy-on-your-car/ [https://perma.cc/5CJC-P9DZ].

5. For a look at the machine learning methods and algorithms behind an ALPR system, including how to build one from the ground up, see Quang Nguyen, Detect and Recognize Vehicle’s License Plate with Machine Learning and Python — Part 1: Detection License Plate with Wpod-Net, MEDIUM (Apr. 11, 2020), https://medium.com/@quangnhatnguyenle/detect-and- recognize-vehicles-license-plate-with-machine-learning-and-python-part-1-detection-795fda47e 922 [https://perma.cc/RRN5-BMGH].

6. Tom Simonite, AI License Plate Readers Are Cheaper—So Drive Carefully, WIRED (Jan. 27, 2020, 8:00 AM), https://www.wired.com/story/ai-license-plate-readers-cheaper-drive- carefully/ [https://perma.cc/WY5V-SG5R] (describing a new “boost from AI . . . that will make the [ALPR] device better at reading plates at high speed or in bad weather.”).

7. Rohrlich, supra note 4. But see Green v. City of San Francisco: (“ALPR”) mistakenly identified Green’s Lexus as a stolen vehicle. Without visually confirming the license plate, Sergeant Kim made a “high-risk” stop during which Green was held at gunpoint by multiple officers, handcuffed, forced to her knees, and detained for up to twenty minutes. She was released only after officers eventually ran her plate and discovered the ALPR mistake and that

8. E.g., Automated License Plate Readers (ALPRs), ELECTRONIC FRONTIER FOUNDATION, https://www.eff.org/pages/automated-license-plate-readers-alpr [https://perma.cc/7QLQ-6DH2] (last updated Aug. 28, 2017).

9. Id.

10. Matt Katz, NYPD Pays $1 Million, Vows Surveillance Reforms After Settling with Muslims in New Jersey, WNYC (Apr. 5, 2018), https://www.wnyc.org/story/nypd-vows- surveillance-reforms-after-settling-lawsuit-muslims-new-jersey/ [https://perma.cc/DC6T-EE ZV].

11. Chris Francescani, NYPD Expands Surveillance Net to Fight Crime as well as Terrorism, REUTERS (June 21, 2013, 11:24 AM), https://www.reuters.com/article/usa-ny- surveillance/nypd-expands-surveillance-net-to-fight-crime-as-well-as-terrorism-idUSL2N0EV0 D220130621 [https://perma.cc/M4EY-NGTK]. More recently, the state of Maryland captured over 500 million scans in 2020 alone. Harwell & Timberg, supra note 1.

12. Adam Goldman & Matt Apuzzo, NYPD Defends Tactics Over Mosque Spying; Records Reveal New Details On Muslim Surveillance, HUFFPOST (Apr. 25, 2012), https://www.huffpost.com/entry/nypd-defends-tactics-over_n_1298997 [https://perma.cc/J4GZ- EHZE].

13. Newcombe, supra note 3.

14. Simonite, supra note 6 (detailing how a small-town PD, near this author’s hometown, is now tapped into several public security cameras because its supplier “charges as little as $50 per month” for each camera).

15. E.g., UCLA Policy 134, UCLA.EDU, http://www.adminpolicies.ucla.edu/APP/Number/ 134.0 [https://perma.cc/36B9-4XXH] (last visited Jan. 11, 2022); Traffic Rules and Regulations, UNIV. S. ALABAMA, https://www.southalabama.edu/departments/parkingservices/rulesandregs .html [https://perma.cc/9BPD-MJ9S] (last visited Jan. 11, 2022).

16. See infra text accompanying notes 19–23.

17. Orin S. Kerr, The Mosaic Theory of the Fourth Amendment, 111 MICH. L. REV. 311, 313 (2012); see also discussion infra Section I.A.

18. Commonwealth v. McCarthy, 142 N.E.3d 1090, 1095 (Mass. 2020).

19. One exception from 2019 concerned holding the data under a specific state statute, Virginia’s Data Act. Neal v. Fairfax Cnty. Police Dep’t, 2019 WL 1438078 at .*2–4 (Va. Cir. 2019).

20. 475 U.S. 106, 114–15 (1986) (finding no expectation privacy when “[t]he exterior of a car . . .is thrust into the public eye”).

21. E.g., Gannett Co., Inc. v. Cnty. of Monroe, 4 N.Y.S.3d 847, 904 (Sup. Ct. 2015) (“Since a vehicle either being driven or parked on the street occurs in public, a recorded ‘read’ by a license plate reader is not an invasion of personal privacy.”).

22. United States v. Wilcox, 415 Fed. Appx. 990, 991 (11th Cir. 2011) (per curiam).

23. The defendant did not object in a timely manner. Id. at 992.

24. Maynard v. United States, 615 F.3d 544 (D.C. Cir. 2010), aff’d sub nom. United States v. Jones, 565 U.S. 400 (2012).

25. Kerr, supra note 17.

26. Paul Rosenzweig, In Defense of the Mosaic Theory, LAWFARE (Nov. 29, 2017, 3:18 PM), https://www.lawfareblog.com/defense-mosaic-theory [https://perma.cc/V4BT-4QFA].

27. Id.; see also Kerr, supra note 17, at 345.

28. Kerr, supra note 17.

29. Jones, 565 U.S. 400, 402 (2012).

30. Id. at 418 (Alito, J., concurring in judgment).

31. Id. at 415 (Sotomayor, J., concurring) (expressing concerns with a “precise, comprehensive record of a person’s public movements that reflects a wealth of detail about her familial, political, professional, religious, and sexual associations” and the fact that “the Government’s unrestrained power to assemble data that reveal private aspects of identity is susceptible to abuse”); id. at 418 (Alito, J., concurring in judgment) (“asking whether respondent’s reasonable expectations of privacy were violated by the long-term monitoring”).

32. E.g., Kerr, supra note 17 (“[C]oncurring opinions signed or joined by five of the justices endorsed some form of the D.C. Circuit’s mosaic theory.”); Christopher Slobogin, Making the Most of United States v. Jones in a Surveillance Society: A Statutory Implementation of Mosaic Theory, 8 DUKE J. CONST. L. & PUB. POL’Y 1, 3 (2012) (same).

33. E.g., Jessica Gutierrez-Alm, Note, The Privacies of Life: Automatic License Plate Recognition is Unconstitutional Under the Mosaic Theory of Fourth Amendment Privacy Law, 38 HAMLINE L. REV. 127 (2015); Rachel Levinson-Waldman, Hiding in Plain Sight: A Fourth

34. 142 N.E.3d 1090, 1095 (Mass. 2020).

35. Id. at 1101 (citing United States v. Knotts, 460 U.S. 276, 283–85 (1983)).

36. Knotts, 460 U.S. at 281.

37. Id. at 284.

38. McCarthy, 142 N.E.3d at 1101. But see discussion infra Section II.A.

39. Id. at 1102–03 (alteration in original) (quoting Katz v. United States, 389 U.S. 347, 351 (1967)).

40. Id.

41. Id. at 1103–04.

42. Id. at 1104 (citations omitted).

43. Id. at 1105 (quoting Dow Chem. Co. v. United States, 476 U.S. 227, 238 n.5 (1986)).

44. Id. at 1106.

45. Id. at 1104.

46. 533 U.S. 27 (2001).

47. McCarthy, 142 N.E.3d at 1102. It should be noted that a similar pronouncement was “a proposition the Court was careful not to announce in Jones,” in the words of Justice Kennedy. Carpenter v. United States, 138 S. Ct. 2206, 2231 (2018) (Kennedy, J., dissenting).

48. Id. at 1104.

49. See United States v. Wilford, 961 F. Supp. 2d 740, 772 (D. Md. 2013) (“[T]he mosaic theory was not adopted as a holding by the Supreme Court.”).

50. Maynard v. United States, 615 F.3d 544 (D.C. Cir. 2010), aff’d sub nom. United States v. Jones, 565 U.S. 400 (2012).

51. Kerr, supra note 17.

52. See, e.g., sources cited supra notes 32–33.

53. McCarthy, 142 N.E.3d at 1101 (quoting Carpenter v. United States, 138 S. Ct. 2206, 2217 (2018)).

54. Carpenter, 138 S. Ct. at 2217 (citing Jones, 565 U.S. at 430 (Alito, J., concurring in judgment), 415 (Sotomayor, J., concurring)).

55. He cited directly to both Jones concurrences, which should clarify the point.

56. Other courts have also used the same partial quote from Carpenter to imply that it stands for more than it does. E.g., United States v. Yang, 958 F.3d 851, 862 (9th Cir. 2020) (Bea, C.J., concurring in judgment).

57. Orin Kerr, Four Thoughts on the Briefing in Carpenter v. United States, LAWFARE (Nov. 17, 2017, 3:06 PM), https://www.lawfareblog.com/four-thoughts-briefing-carpenter-v-united- states [https://perma.cc/N65J-5EJX].

58. Carpenter, 138 S. Ct. at 2244 n.10 (Thomas, J., dissenting).

59. Id. at 2211 (majority opinion).

60. Id. at 2218.

61. United States v. Jones, 565 U.S. 400, 415 (2012) (Sotomayor, J., concurring) (quoting People v. Weaver, 909 N.E.2d 1195, 1199 (N.Y. 2009)).

62. United States v. Yang, 958 F.3d 851, 862 (9th Cir. 2020) (Bea, C.J., concurring in judgment).

63. Id. (quoting Carpenter, 138 S. Ct. at 2217).

64. Id. (quoting Carpenter, 138 S. Ct. at 2217).

65. See Katz v. United States, 389 U.S. 347, 351 (1967) (citations omitted) (“What a person knowingly exposes to the public . . . is not a subject of Fourth Amendment protection. But what he seeks to preserve as private, even in an area accessible to the public, may be constitutionally protected.”).

66. Carpenter, 138 S. Ct. at 2218 (quoting Riley v. California, 573 U.S. 373, 385 (2014)) (cleaned up).

67. Id. (quoting Caldwell v. Lewis, 417 U.S. 583, 590 (1974) (plurality opinion)).

68. Id.

69. Four members of the Carpenter majority and three dissenters make up the current Court, and Justice Kavanaugh voted against Jones at the circuit level. United States v. Jones, 656 F.3d 766, 769 (D.C. Cir. 2010). It should also be noted that there was an additional factor in Carpenter, not present with ALPR, that cut against privacy, which is that the government sought the CSLI from a third party, the cell service provider. Carpenter, 138 S. Ct. at 2212.

70. The McCarthy court even admitted that “no ALPR network is likely to be as detailed in its surveillance as GPS or CSLI data.” Commonwealth v. McCarthy, 142 N.E.3d 1090, 1104 (Mass. 2020); see also id. at 1102 (quoting Carpenter, 138 S. Ct. at 2218) (noting how CSLI “achieves ‘near perfect surveillance’”).

71. Id. (quoting Carpenter, 138 S. Ct. at 2217). When Carpenter was remanded, the Sixth Circuit stated: “Key to the Court’s reasoning was the inability of CSLI to distinguish between public and private life.” United States vs Carpenter, 926 F.3d 313, 316 (6th Cir. 2019).

72. McCarthy, 142 N.E.3d at 1101 (Mass. 2020) (quoting New York v. Class, 475 U.S. 106, 114 (1986)); see also cases cited supra notes 19–23, infra notes 98–99, and accompanying text.

73. Katz v. United States, 389 U.S. 347, 351 (1967).

74. McCarthy, 142 N.E.3d at 1101 (citing United States v. Knotts, 460 U.S. 276, 285 (1983) and Commonwealth v. Starr, 773 N.E.2d 981, 984–85 (Mass. App. Ct. 2002)).

75. Id. at 1103.

76. See supra notes 50–52 and accompanying text.

77. See supra notes 53–58 and accompanying text.

78. McCarthy, 142 N.E.3d at 1103.

79. 468 U.S. 705, 707 (1984).

80. Id. at 714.

81. Id.

82. Id. at 715 (quoting United States v. Knotts, 460 U.S. 276, 281 (1983)).

83. Id.

84. See discussion supra Section I.B.

85. Katz v. United States, 389 U.S. 347, 351 (1967).

86. Id. at 361 (Harlan, J., concurring).

87. E.g., United States v. Jones, 565 U.S. 400, 411 (2012) (citations omitted) (“[A]n open field, unlike the curtilage of a home, is not one of those protected areas enumerated in the Fourth Amendment.”).

88. 846 F. Supp. 2d 384, 404 (D. Md. 2012).

89. Carpenter v. United States, 138 S. Ct. 2206, 2220 (2018).

90. Id. at 2215 (quoting United States v. Knotts, 460 U.S. 276, 284 (1983)).

91. Commonwealth v. McCarthy, 142 N.E.3d 1090, 1104–05 (Mass. 2020).

92. United States v. Yang, 958 F.3d 851, 853 (9th Cir. 2020).

93. See Graham, 846 F. Supp. 2d at 404 (“[Knotts and Karo] stand for the proposition that law enforcement conducts a Fourth Amendment ‘search’ when it utilizes tracking technology that allows surveillance in locations that police could not monitor in the absence of that technology.”).

94. McCarthy, 142 N.E.3d at 1098 (quoting Commonwealth v. Almonor, 120 N.E.3d 1183, 1191 (2019)).

95. Id. (citing Kyllo, 533 U.S. 27, 34 (2001)).

96. Kyllo, 533 U.S. at 29.

97. Id. at 40.

98. Id. at 32 (citing Dow Chem., 476 U.S. 227, 234–35 (1986)).

99. Id. at 34.

100. Id. at 40; cf. United States v. Knotts, 460 U.S. 276, 282 (1983) (reasoning that the radio technology “merely augmented officers’ physical abilities and did not provide more information than officers could have obtained by visual surveillance”).

101. Kyllo, 533 U.S. at 40.

102. Cyrus Farivar, New Software Watches for License Plates, Turning You into Little Brother, ARS TECHNICA (Dec. 5, 2015, 12:30 PM), https://arstechnica.com/information- technology/2015/12/new-open-source-license-plate-reader-software-lets-you-make-your-own- hot-list/ [https://perma.cc/RET3-4LUN]; see also Rohrlich, supra note 4.

103. Ella Fassler, Neighborhood Watch Has a New Tool: Privately Owned License-Plate Readers, MEDIUM: ONEZERO (Nov. 12, 2020), https://onezero.medium.com/neighborhood-watch- has-a-new-tool-privately-owned-license-plate-readers-302f296abb27 [https://perma.cc/MY4W- NXLH].

104. Josh Kaplan, License Plate Readers Are Creeping into Neighborhoods Across the Country, SLATE: FUTURE TENSE (July 10, 2019, 7:30 AM), https://slate.com/technology/2019/ 07/automatic-license-plate-readers-hoa-police-openalpr.html [https://perma.cc/35W9-5JQY].

105. Mike Petridis, In General Public Use: An Unnecessary Test to Determine Whether the Use of Advanced Sensing Technology Was a Fourth Amendment Search, TOURO L. REV. BLOG (Apr. 21, 2020), https://tourolawreviewblog.wordpress.com/2020/04/21/in-general-public-use-an -unnecessary-test-to-determine-whether-the-use-of-advanced-sensing-technology-was-a-fourth- amendment-search/ [https://perma.cc/WX2J-R7RJ].

106. Kyllo, 533 U.S. at 36 (“[T]he rule we adopt must take account of more sophisticated systems that are already in use or in development.”).

107. Florida v. Jimeno, 500 U.S. 248, 250 (1991) (citing Katz v. United States, 389 U.S. 347, 360 (1967) (Harlan, J., concurring)); see also sources cited supra note 58.

108. See Katz, 389 U.S. at 348 (majority opinion).

109. Kyllo, 533 U.S. at 29.

110. Id. at 31 (quoting Silverman v. United States, 365 U.S. 505, 511 (1961)).

111. Id. (emphasis added) (citations omitted). The Court later observed that the inside of a home is “the prototypical and hence most commonly litigated area of protected privacy.” Id. at 34.

112. Id. at 33 (emphasis in original) (citing Dow Chem. Co. v. United States, 476 U.S. 227, 237 n.4 (1986)).

113. See id. at 32 (quoting California v. Ciraolo, 476 U.S. 207, 213 (1986)) (“[T]he Fourth Amendment protection of the home has never been extended to require law enforcement officers to shield their eyes when passing by a home on public thoroughfares.”).

114. Id. at 40.

115. Id. at 34.

116. New York v. Class, 475 U.S. 106, 114 (1986). Many states retain ownership of the plate, further removing any privacy interest. E.g., License Plates & Registration, FLA. HIGHWAY

117. Commonwealth v. McCarthy, 142 N.E.3d 1090, 1104 (Mass. 2020) (quoting Carpenter v. United States, 138 S. Ct. 2206, 2218 (2018)).

118. Id. at 1102 (emphasis in original) (quoting Commonwealth v. Augustine, 4 N.E.3d 846, 863–64 (Mass. 2014)).

119. See cases cited supra note 100 and accompanying text; see also Jeff Weiner, UCF Scanning License Plates of Cars on Campus to Check Against Police Databases, ORLANDO SENTINEL (June 20, 2019, 4:09 PM), https://www.orlandosentinel.com/news/crime/os-ne-ucf- license-plate-scanners-on-campus-parking-20190620-pop76kgusbfrdaw6iz2our2rmm-story.html [https://perma.cc/9PGQ-EFDC] (university police chief calling its new ALPR network a “force multiplier”).

120. McCarthy, 142 N.E.3d at 1106.

121. See, e.g., United States v. Knotts, 460 U.S. 276, 285 (1983) (“A police car following [the driver] at a distance could have observed him.”).

122. See People v. Weaver, 909 N.E.2d 1195, 1199 (N.Y. 2009) (“GPS is not a mere enhancement of human sensory capacity, it facilitates a new technological perception of the world.”).

123. Knotts, 460 U.S. at 282.

124. Kyllo v. United States, 533 U.S. 27, 32 (2001) (citation omitted).

125. See United States v. Jones, 565 U.S. 400, 416 (Sotomayor, J., concurring); id. at 420 (Alito, J., concurring in judgment).

126. This includes the starting point for McCarthy’s discussion section, for example. Commonwealth v. McCarthy, 142 N.E.3d 1090, 1097 (Mass. 2020).

127. 389 U.S. 347 (1967).

128. Carpenter v. United States, 138 S. Ct. 2208, 2213 (2018) (quoting Smith v. Maryland, 442 U.S. 735, 740 (1979)).

129. Jones, 565 U.S. at 416 (Sotomayor, J., concurring) (“I would ask whether people reasonably expect that their movements will be recorded and aggregated.”); id. at 419 (Alito, J., concurring in judgment) (“I would analyze the question presented in this case by asking whether respondent’s reasonable expectations of privacy were violated by the long-term monitoring.”).

130. This includes the McCarthy court. See supra note 70.

131. Gutierrez-Alm, supra note 33, at 151–52; see also supra note 70 and accompanying text.

132. See supra note 61 and accompanying text.

133. As the McCarthy court acknowledged, “[i]t is an entirely ordinary experience to drive past a police officer in a cruiser observing traffic.” Commonwealth v. McCarthy, 142 N.E.3d 1090, 1106 (Mass. 2020) (citation omitted).

134. See supra note 13 and accompanying text.

135. See Jones, 565 U.S. at 430 (Alito, J., concurring in judgment) (“ask[ing] whether the use of GPS tracking . . . involved a degree of intrusion that a reasonable person would not have anticipated”). But see Gutierrez-Alm, supra note 33, at 152 (arguing for the use of Professor Kerr’s probabilistic model to overcome this when it comes to aggregation).

136. 142 N.E.3d at 1105.

137. Id. at n.14.

138. Shawn Musgrave, Boston Police Halt License Plate Scanning Program, BOSTON GLOBE (Dec. 14, 2013), https://www.bostonglobe.com/metro/2013/12/14/boston-police-suspend-use- high-tech-licence-plate-readers-amid-privacy-concerns/B2hy9UIzC7KzebnGyQ0JNM/story.html [https://perma.cc/6HPK-M34X]. The program resumed in 2018. Shawn Musgrave, Boston Police Resume Using License Plate Readers After Accidental Release of Data, BOSTON GLOBE (May 6, 2018), http://www0.bostonglobe.com/metro/2018/05/06/boston-police-resume-using-license- plate-readers-after-accidental-release-data/gZrC8ozxad9GxcymIxtLfO/story.html [https://perma .cc/8687-AZJ5].

139. 476 U.S. 207, 209 (1986).

140. Id.; see also discussion supra Section II.C.

141. Ciraolo, 476 U.S. at 211.

142. Id. at 213–14.

143. Id. at 214 (citing Katz v. United States, 389 U.S. 347, 362 (1967) (Harlan, J., concurring)).

144. 488 U.S. 445 (1989).

145. Id. at 449–51 (plurality opinion) (citing Ciraolo, 476 U.S. at 215).

146. Id. at 450–51 (emphasis added). Justice Blackmun, in his dissent, agreed that “the reasonableness of Riley’s expectation depends, in large measure, on the frequency of nonpolice helicopter flights at an altitude of 400 feet.” Id. at 467 (Blackmun, J., dissenting).

147. See supra note 13 and accompanying text.

148. See, e.g., sources cited supra notes 15, 119.

149. U.S. CONST. amend. IV (emphasis added).

150. United States v. Jones, 565 U.S. 400, 413 (2012).

151. 21 N.Y.3d 515, 518 (2013).

152. People v. Weaver, 909 N.E.2d 1195 (N.Y. 2009); see supra note 61 and accompanying text.

153. Cunningham, 21 N.Y.3d at 522.

154. Id. at 523 (“Where an employer conducts a GPS search without making a reasonable effort to avoid tracking an employee outside of business hours, the search . . . must be considered unreasonable.”).

155. See text accompanying supra notes 91–92.

156. Commonwealth v. McCarthy, 142 N.E.3d 1090, 1095 (Mass. 2020).

157. E.g., United States v. Yang, 958 F.3d 851, 853 (9th Cir. 2020).

158. Exclusionary Rule, LEGAL INFO. INST., https://www.law.cornell.edu/wex/exclusionary _rule [https://perma.cc/5MDP-5TNC] (last visited Jan. 7, 2022).

159. 428 U.S. 465, 486 (1976).

160. Id. at 485; see also Elkins v. United States, 364 U.S. 206, 217 (1960) (“The [exclusionary] rule is calculated to prevent, not to repair.”).

161. Jones v. United States, 362 U.S. 257, 261 (1960) (“Ordinarily, then, it is entirely proper to require of one who seeks to challenge the legality of a search as the basis for suppressing relevant evidence . . . if the allegation be disputed that he establish, that he himself was the victim of an invasion of privacy.”).

162. Br. for Defendant-Appellant, 2019 WL 3996632, at .*19.

163. See Herring v. United States, 555 U.S. 135, 148 (2009) (“[T]he deterrent effect of suppression must be substantial and outweigh any harm to the justice system.”).

164. See generally Good Faith Exception to Exclusionary Rule, LEGAL INFO. INST., https://www.law.cornell.edu/wex/good_faith_exception_to_exclusionary_rule [https://perma.cc/ FJ5Z-9339] (last visited Jan. 8, 2022).

165. Id. (“In Davis v. U.S., the U.S. Supreme Court ruled that the exclusionary rule does not apply when the police conduct a search in reliance on binding appellate precedent allowing the search.”).

166. McCarthy, 142 N.E.3d at 1095–96.

167. Id. at 1096.

168. Id. at 1101 (citing Commonwealth v. Starr, 773 N.E.2d 981, 984–85 (Mass. App. Ct. 2002)).

169. See Appellee’s Br., 2019 WL 4134166; cf. Tracey v. State, 152 So. 3d 504, 526 (Fla. 2014) (ruling that the good faith exception did not apply).

170. 846 F. Supp. 2d 384, 405 (D. Md. 2012).

171. Id. at 405–06.

172. Id.

173. Id. at 406.

174. United States vs Carpenter, 926 F.3d 313, 314 (6th Cir. 2019).

175. 564 U.S. 229, 232 (2011).

176. McCarthy, 142 N.E.3d at 1104.

177. The court admitted as much when it said that “we cannot say precisely how detailed a picture of the defendant’s movements must be revealed to invoke constitutional protections.” Id. at 1106.

178. See supra note 168 and accompanying text.

179. For similar rationale concerning qualified immunity, see, e.g., Karen M. Blum, The Qualified Immunity Defense: What’s “Clearly Established” and What’s Not, 24 TUORO L. REV. 4 (2014).

180. Exclusionary Rule, supra note 158.

181. 467 U.S. 431, 443 (1984).

182. See sources cited supra notes 103–04 and accompanying text.

183. But see United States v. Jones, 565 U.S. 400, 416 (2012) (Sotomayor, J., concurring) (citing Kyllo v. United States, 533 U.S. 27, 35 n.2 (2001)) (“I do not regard as dispositive the fact that the government might obtain the fruits of GPS monitoring through lawful conventional surveillance techniques.”).

184. Fassler, supra note 103.

185. McCarthy, 142 N.E.3d at 1104–06; see also discussion infra Section III.B.

186. Law enforcement in a state with an ALPR statute (which Massachusetts is not, see infra note 217) may have additional protection under Illinois v. Krull, 480 U.S. 340, 355 (1987). See Kerr, supra note 17, at 341–42; United States v. Graham, 846 F. Supp. 384, 405–06 (D. Md. 2012); United States v. Carpenter, 926 F.3d 313, 317–18 (6th Cir. 2019).

187. See Gutierrez-Alm, supra note 33, at 152–55 (recognizing established precedent of a lack of an expectation of privacy in public, noting that the mosaic theory “must be employed in the consideration of ALPR data in order to demonstrate a reasonable expectation of privacy,” and arguing that it should be adopted).

188. See United States v. Jones, 656 F.3d 766, 769 (D.C. Cir. 2010) (Sentelle, C.J., dissenting) (“The sum of an infinite number of zero-value parts is also zero.”). It should be noted that Justice Kavanaugh joined that opinion. For the opposing view, see Rosenzweig, supra note 26 (“with enough data 1+1+1 really does equal 17”).

189. Graham, 846 F. Supp. at 401; accord State v. Muhammad, 451 P.3d 1060, 1072–73 (Wash. 2019) (discussing “practical problems inherent in this [mosaic] theory”); see also infra notes 202–03 and accompanying text.

190. Graham, 846 F. Supp. 2d at 401 (citing Orin S. Kerr, D.C. Circuit Introduces “Mosaic Theory” of Fourth Amendment, Holds GPS Monitoring a Fourth Amendment Search, THE VOLOKH CONSPIRACY (Aug. 6, 2010, 2:46 PM), http://volokh.com/2010/08/06/d-c-circuit- introduces-mosaic-theory-of-fourth-amendment-holds-gps-monitoring-a-fourth-amendment- search/ [https://perma.cc/9EUB-FSXV]).

191. Albeit perhaps with more cameras. Commonwealth v. McCarthy, 142 N.E.3d 1090, 1104 (Mass. 2020) (“With enough cameras in enough locations . . .”).

192. See Graham, 846 F. Supp. 2d at 402 (quoting Kerr, supra note 190) (“that first day of monitoring eventually and retroactively becomes unconstitutional. It becomes part of the mosaic, and the key point of Maynard is that the entire mosaic is considered one entity.”).

193. Id. at 401. Granted, Graham was decided before Carpenter—but see text accompanying supra notes 133–34.

194. See Kerr, supra note 17; Kerr, supra note 121; Kerr, supra note 190.

195. Kerr, supra note 57.

196. Id.

197. See id.; see generally Terry Stop and Frisk: Doctrine and Practice, CONSTITUTION ANNOTATED, https://constitution.congress.gov/browse/essay/amdt4_4_4_1_1/ [https://perma.cc/ 3VDP-346Z] (last visited Jan. 11, 2022).

198. Oliver v. United States, 466 U.S. 170, 181 (1984) (collecting cases).

199. Id. at 181–82 (citation omitted).

200. Kerr, supra note 57; see also discussion infra Section IV.A.

201. See Orin S. Kerr, Automated License Plate Readers, the Mosaic Theory, and the Fourth Amendment, REASON: THE VOLOKH CONSPIRACY (Apr. 22, 2020, 5:46 AM), https://reason.com/volokh/2020/04/22/automated-license-plate-readers-the-mosaic-theory-and- the-fourth-amendment/ [https://perma.cc/SLJ3-VXYR].

202. 152 So. 3d 504, 520 (Fla. 2014).

203. Id. (citing Oliver, 466 U.S. at 181).

204. Id. at 521 (citations omitted); see also Wyoming v. Houghton, 526 U.S. 295, 306 (1999) (finding that “practical realities . . . militate in favor of the needs of law enforcement”).

205. Tracey, 152 So. 3d at 521–22.

206. Compare id. at 525–26, with Carpenter v. United States, 138 S. Ct. 2206, 2220 (2018).

207. E.g., Paul Ohm, The Many Revolutions of Carpenter, 32 HARV. J. L. & TECH. 357, 393 (2019) (concluding that ALPR usage “is likely to be a very close call” after Carpenter).

208. United States v. Graham, 846 F. Supp. 2d 384, 401 (D. Md. 2012); see also discussion supra Section III.A.

209. Carpenter, 138 S. Ct. at 2212.

210. Id. at 2221.

211. Records requests by a team of researchers found 2.5 billion scans among 173 responding agencies over 2016–17. Each agency shared directly with an average of 160 others and upwards of 850. One denied participating but eventually admitted that while it did not have any scanners of its own, it could access over 500 other agencies’. Dave Maass & Beryl Lipton, Data Driven: Explore How Cops Are Collecting and Sharing Our Travel Patterns Using Automated License Plate Readers, MUCKROCK (Nov. 15, 2018), https://www.muckrock.com/ news/archives/2018/nov/15/alpr-landing-page/ [https://perma.cc/MVF9-K4RN].

212. The Graham court took judicial notice that Professor Kerr is a leading scholar in this area, testifying before Congress on privacy issues, and that one of his articles was cited by both the majority opinion and a concurrence in Jones. Graham, 846 F. Supp. 2d at 402 n.14.

213. Kerr, supra note 201.

214. The fact that in-court objections to ALPRs often come on motions to suppress, as in McCarthy, is indicative of their effectiveness. Commonwealth v. McCarthy, 142 N.E.3d 1090, 1095 (Mass. 2020).

215. United States v. Jones, 565 U.S. 400, 429 (Alito, J., concurring in judgment).

216. E.g., Lisa Soronen, Justice Scalia’s Impact on State and Local Government, NAT’L CONF. STATE LEGISLATURES (Feb. 15, 2016), https://www.ncsl.org/blog/2016/02/15/justice- scalias-impact-on-state-and-local-government.aspx [https://perma.cc/H6G8-ACEA].

217. Arkansas, California, Colorado, Florida, Georgia, Maine, Maryland, Minnesota, Montana, Nebraska, New Hampshire, North Carolina, Oklahoma, Tennessee, Utah, and Vermont. Automated License Plate Readers: State Statutes, NAT’L CONF. STATE LEGISLATURES, https://www.ncsl.org/research/telecommunications-and-information-technology/state-statutes- regulating-the-use-of-automated-license-plate-readers-alpr-or-alpr-data.aspx [https://perma.cc/ BHU4-4G74] (last updated Apr. 9, 2021).

218. Id.

219. Sidney Fussell, One Clear Message From Voters This Election? More Privacy, WIRED (Nov. 4, 2020, 8:26 PM), https://www.wired.com/story/one-clear-message-voters-election-more- privacy [https://perma.cc/RR5T-LUEJ].

220. GA. CODE ANN. § 35-1-22(b) (2021).

221. N.H. REV. STAT. ANN. § 261:75-b, VIII (2021).

222. Clara Turnage, High-Tech Devices Help Police in Little Rock, but Some Say Tools such as License Plate Readers Raise Thorny Issues, ARKANSAS DEMOCRAT GAZETTE (Apr. 1, 2019, 4:30 AM), https://www.arkansasonline.com/news/2019/apr/01/high-tech-devices-help-police-in- lr-201/ [https://www.arkansasonline.com/news/2019/apr/01/high-tech-devices-help-police-in-lr- 201/].

223. E.g., Ryan Rodenberg, United States of Sports Betting: An Updated Map of Where Every State Stands, ESPN: CHALK (Nov. 3, 2020), https://www.espn.com/chalk/story/_/id/197 40480/the-united-states-sports-betting-where-all-50-states-stand-legalization [https://perma.cc/ H2DK-SL46].

224. S.B. 0243, 101st Gen. Assemb. (Ill. 2019); A.B. A7254, 2019–2020 Leg. Sess. (N.Y. 2020).

225. H.B. 3141, 191st Gen. Court (Mass. 2020).

226. United States v. Graham, 846 F. Supp. 2d 384, 396 (D. Md. 2012); 18 U.S.C. §§ 2701– 12.

227. United States v. Carpenter, 926 F.3d 313, 317 (6th Cir. 2019).

228. 18 U.S.C. § 2703(d).

229. See supra note 8 and accompanying text.

230. 18 U.S.C. § 2703(d).

231. MINN. STAT. § 13.824 (2020); Minnesota State Patrol Audit, MINNESOTA LEGISLATURE (Nov. 27, 2017), https://www.leg.mn.gov/docs/2020/mandated/200650.pdf [https://perma.cc/ F8GP-567A].

232. See supra text accompanying notes 10–12.

233. Francescani, supra note 11.

234. E.g., Cyrus Farivar, Cops Are Freaked Out That Congress May Impose License Plate Reader Limits, ARS TECHNICA (Mar. 15, 2015, 11:00 AM), https://arstechnica.com/tech- policy/2015/03/cops-are-freaked-out-that-congress-may-impose-license-plate-reader-limits/ [https://perma.cc/8Z2B-VT62].

235. See Francescani, supra note 11 (“There [was] no outside monitoring of this system at all.”).

236. Ángel Díaz, A Bill to Oversee 21st Century Police Surveillance, BRENNAN CTR. FOR JUST. (Feb. 12, 2020), https://www.brennancenter.org/our-work/analysis-opinion/bill-oversee- 21st-century-police-surveillance [https://perma.cc/9ML5-JL3W].

237. The Public Oversight of Surveillance Technology (POST) Act: A Resource Page, BRENNAN CTR. FOR JUST., https://www.brennancenter.org/our-work/research-reports/public- oversight-surveillance-technology-post-act-resource-page [https://perma.cc/P7SF-96X5] (last updated Mar. 5, 2021).

238. Id.; see Technology & Equipment – NYPD, NYC: NYPD, https://www1.nyc.gov/site/ nypd/about/about-nypd/equipment-and-tools.page [https://perma.cc/3ZSH-YFFH] (last visited Jan. 11, 2022).

239. Díaz, supra note 236.

240. Ed Drantch, You’re Being Recorded: Millions of License Plates Tracked with Automatic Plate Readers in Buffalo, WKBW BUFFALO (Nov. 19, 2020, 11:27 PM), https://www.wkbw.com/news/i-team/youre-being-recorded-millions-of-license-plates-tracked- with-automatic-plate-readers-in-buffalo [https://perma.cc/4YQR-2GHC].

241. Id.

242. Laura Bronner, Why Statistics Don’t Capture the Full Extent of the Systemic Bias In Policing, FIVETHIRTYEIGHT (June 25, 2020), https://fivethirtyeight.com/features/why-statistics- dont-capture-the-full-extent-of-the-systemic-bias-in-policing/ [https://perma.cc/U78B-NS9W].

243. Id.; see also Andrew D. Selbst, Disparate Impact in Big Data Policing, 52 GA. L. REV. 109 (2017).

244. Simonite, supra note 6.

245. Bronner, supra note 242.

246. See supra text accompanying note 13.

247. See Simonite, supra note 6 (“California’s Highway Patrol must delete ALPR data after 60 days unless it is being used as evidence of a felony.”); H.B. 3141, 191st Gen. Court (Mass. 2020).

248. See Julia Coin, License Plate Readers Installed in UF Area, THE GAINESVILLE SUN (Nov. 28, 2020, 5:30 AM), https://www.gainesville.com/story/news/local/2020/11/28/license- plate-readers-installed-near-university-florida-campus/6423509002/ [https://perma.cc/QM6T- F6XH] (“Officers can’t anonymously run tags and look for exes”); cf. Simonite, supra note 6 (noting that “Los Angeles law enforcement agencies made tens of thousands of license plate queries each year”).

249. Cf. sources cited supra note 6. Section 3(b) of the proposed Massachusetts bill would require this. H.B. 3141, 191st Gen. Court (Mass. 2020).

250. E.g., Weiner, supra note 119.

251. Id.

252. Heather Landers, Park Safely in the University Central Fla Parking Structures, CHANGE.ORG, https://www.change.org/p/university-central-florida-park-safely-in-the-university- central-fla-parking-structures [https://perma.cc/6USL-HYYD] (last visited Jan. 7, 2022).

253. Id.

254. Which States Require a Front License Plate?, AUTOLIST (May 11, 2020), https://www.autolist.com/guides/front-license-plate [https://perma.cc/E7QX-J6TL].

255. Coincidentally, the author’s university announced the installation of readers in late 2020. Coin, supra note 248.