Borders and bodies are increasingly regulated by data-capturing mechanisms spread across the world through information and communication technologies. This article traces the features and implications of such a border-body datalogical entanglement through the figure of the drug mule. It analyzes government documents and recorded case studies to argue that this figure emerges from an assemblage of cultural narratives, legal structures, human labor, technical practices, and biological processes. The datalogical drug mule is already implicated in a struggle over what, and how, data is meaningful and actionable. Investigating this figure allows us to begin disentangling the data-driven mechanisms that constitute modern borders and bodies while at the same time accounting for analog continuities in contemporary practices of border security.
Rosa Montoya de Hernandez swallowed eighty-eight balloons filled with cocaine, boarded Avianca flight 80 in Bogotá, Colombia, and arrived at Los Angeles International Airport after midnight on March 5, 1983. She spoke little English, had no family or friends in the United States, and carried five thousand dollars in cash, which she claimed was for buying merchandise for her husband's store. A customs inspector detained her because she fit the profile of a “swallower,” or drug mule, and placed her in an observation room overnight.1 She was subjected to a pat down and strip search but she refused to consent to an X-ray examination. After sixteen hours in detention, she had yet to defecate, so customs agents obtained a court order allowing a physician to conduct a pregnancy test, an X-ray, and a rectal examination. The physician discovered one cocaine-filled balloon during the examination and Montoya de Hernandez was arrested. Over the next four days, she passed the remaining eighty-seven.
United States v. Montoya de Hernandez is an exemplary case in legal analyses of the so-called war on drugs. The district court convicted Montoya de Hernandez of possession with intent to distribute and importation of cocaine. The United States Court of Appeals for the Ninth Circuit reversed this decision, arguing that the sixteen-hour detention without “clear indication” of smuggling was a violation of her Fourth Amendment rights. The Supreme Court then reversed the appeals decision, holding that customs agents were justified in an extended detention given their reasonable suspicion.2 The case and its aftermath strike at the heart of issues surrounding the Fourth Amendment, the increasing powers of the US border patrol, and the spatial (and ethical) reach of the border. Ultimately, however, United States v. Montoya de Hernandez is not about Rosa; instead, as one immigration lawyer argues, it is “about what the government can subject a [body] to at the border.”3 It is about what a drug mule can tell us about making borders and marking bodies in a technologically interconnected and interdependent world.
Although the legal implications of this case continue to resonate with contemporary geopolitical bordering practices, it belongs to a prior historical moment in terms of the practical dimensions of border management. Recent developments in the technologies of border administration and bodily regulation extend the reach of borders and the control of bodies. For instance, nowadays, passengers in flight manifestos are prescreened before they check in at their departure airport and after they arrive at their destination. Travelers must undergo body screenings before boarding a plane and, sometimes, when going through immigration checkpoints. While drug smuggling through swallowing continues to be common practice, it is a practice that now exists within a data-driven network of transnational flows and blockages. In short, within contemporary narcotrafficking, the drug mule is always already datalogical.
Narcotrafficking, as theorized in this article, is not reducible to drug trafficking. It also encompasses other drug cartel practices, government corruption, policy decisions, surveillance technologies, cultural narratives, military strategies, legal disputes, fragmented bodies, and much more. Narcotrafficking functions as an assemblage, a heterogeneous meshwork comprised of human and nonhuman agents. Conceptualizing the phenomenon in this way avoids reducing it to an economic or political issue by considering other aspects, including the technological and the biological. Likewise, this framing replaces questions of representation with those of mediated construction. Understanding narcotrafficking as a mediated assemblage requires attending to what Karen Barad has described as “things-in-phenomena,” those that emerge from and in turn constitute a phenomenon.4 Surveillance technologies, policy changes, and state-mandated screenings, for instance, are not developed only to fight narcotrafficking. They are also inextricable elements that shape how this phenomenon is apprehended, managed, and lived.
The drug mule emerges as a salient material-semiotic actor within this techno-mediated assemblage. As a gendered and raced figure, the drug mule exposes how newfound data-driven bordering and body-tracking mechanisms import long-standing biases regarding who gets coded as criminal and in what ways. This article's analysis of historical and contemporary women smugglers reveals how, in practice, these biases remain ingrained in the procedural logics of US Customs and Border Protection (CBP). The development of biometric tracking enabled by information and communication technologies (ICTs) represents a continuation of persistent practices of surveilling marked bodies for suspected illegal activity. Yet, because of her criminal status, the drug mule is not only subject to these surveillance networks, but also an active, creative challenger of their mechanisms. Mule recruiters are aware of the characteristics that CBP tracks in incoming travelers and proactively recruit against these. Mules themselves train to convincingly perform their data-selves in order to avoid suspicion and capture by authorities. Drug mules thus become implicated in a struggle over what, and how, data is meaningful and actionable at every point of possible capture.
What follows is an analytical and speculative investigation into the datalogical drug mule as a figuration within the phenomenon of narcotrafficking. At stake is tracing the corporealization of the datalogical drug mule, or how techno-mediated interactions between human and nonhuman agents produce specific bodies and, in turn, how these bodies become coded as drug mules. These techno-mediated interactions carry within them biases about the gendered, raced, and abled aspects of the drug mule, biases that are sometimes at odds with each other. Indeed, throughout this investigation the drug mule shall be referred to with female pronouns. If the corporealization of the drug mule involves the circulation of narratives, the aim is to delve into these racially and gender-coded narratives in order to, simultaneously, reveal them as partial and contested. In other words, the aim is not to elucidate the factual inaccuracy of assuming most drug mules are women but to illustrate how these assumptions have real implications regarding which bodies are allowed to circulate within a techno-mediated assemblage.
As Donna Haraway argues, corporealization is deeply contingent and historical.5 Tracing the emergence of the datalogical drug mule therefore requires accounting for historical differences as well as continuities. The first half of this article analyzes the implications of the rise of digital technologies for contemporary borders, but the second half argues for the persistence of erstwhile non-digital procedures in border management practices. The article holds both arguments together by relying on an expansive definition of algorithms. Algorithms, as theorized in what follows, are logic-driven step-by-step procedures through which data becomes legible and actionable. For instance, recently implemented computational algorithms codify the drug mule by searching for particular demographic characteristics in databases of international travelers. Yet non-computational algorithms such as CBP rules and procedures remain integral to sorting travelers based on visual and behavioral markers. The logics behind both of these are ultimately not that different. The analysis of the datalogical drug mule thus follows on the work of Simone Browne to argue that surveillance practices amplified by digital technologies have longer histories in analog control strategies, many of which remain in place nowadays.6
Finally, this article uses government documents and recorded case studies of drug mules to argue for the dual computational-biological dimension of the data-body and for its implications within the increasingly data-driven security apparatus of the state. It argues that the turn to digital technologies in border management and body imaging gives rise to a border-body datalogical entanglement. The drug mule exploits this entanglement's algorithmic logic, thereby exposing openings for the tactical undoing of data-driven assemblages. This article posits the “return of the biological” as an instance where swallowers further contest the presumption of data's “given-ness”—that is, that it can be apprehended without a procedural method of capture. The CBP's dependence on a non-computational algorithm to interpret a potential drug mule's biological processes signals the importance of attending to both automated and human-dependent mechanisms of capture to comprehend the ontology of data. The datalogical drug mule thus reveals the techno-mediated dimension of contemporary narcotrafficking and, ultimately, illustrates the indissoluble link between the computational and the biological in our contemporary data-driven moment.
THE DRUG MULE's TRAVELS
Historically, the mule has always been a key figure in the drug trade. For most of the twentieth century, drugs were one among many items smuggled into the United States on a regular basis. Smuggling functioned as an entrepreneurial activity that took advantage of the growing intensification of the country's border building and prohibitionist efforts.7 In the wake of the Great Depression, unemployed and otherwise displaced people pursued transnational smuggling of high-value items, such as narcotics, as a means of subsistence.8 Maria Wendt, for instance, arrived on the Heiyo Maru in San Pedro, California, on August 5, 1936, with fifty-four pounds of heroin. Although initially captured by customs officials, she famously escaped, flew to New York, and almost fled the country before the Coast Guard intercepted the ship that would have taken her to Europe. She was of Chinese and German descent, and her ethnic ambiguity allowed both the US and the Mexican authorities to publicly blame various groups for the growing influx of narcotics to North America: Germans, Poles, Jews, Japanese, and Chinese.9 Tying the drug mule's practices to her assumed national origins perpetuated stereotypes about the deviancy or victimhood of these groups as well. The drug mule, as a modern concept, is thus already marked in terms of gender, class, and race.
The mule who transports drugs inside her body, also called a “swallower” or “body packer,” represents a more recent development. The first recorded cases of swallowers in the Americas were in the mid-1970s.10 The creation of customs canine units in the United States in 1986, which limited the ease of smuggling through concealed packages, contributed to a sharp increase in swallowing.11 Customs officials referred to this type of drug trafficking as “the bane of airport security” by the late 1980s.12 In 1995, US customs officials reported more than one hundred arrests for drug swallowing at Miami International Airport, but estimated that at least six drug mules were getting through for every one caught.13 Concomitant with the increase in this type of smuggling, traffickers refined the methods for making the drug pellets that swallowers ingest. The typical packaging consisting of handmade double-bagged condoms gave way to sturdy, machine-pressed pellets covered in hard wax.14 As the methods for body packing became more technically sophisticated, so did the need for tools to detect drug mules.
In writings about narcotrafficking, the smuggler connotes someone who “moves goods for her own enrichment or benefit,” while the mule is “merely a vessel for transportation controlled by others.”15 Questions of agency and medium are central to determining who counts as a mule, and under what circumstances. Traditional narratives code the smuggler as a masculine role and the mule as a feminine one, perpetuating notions about, on one hand, female bodies as a medium of exchange between men and, on the other, femininity as passive and receptive.16 Contemporary sociological and anthropological research reveals that this assumption is unfounded because both men and women work as mules, and in some markets the men outnumber the women.17 Media portrayals, however, propagate this gender asymmetry, and some posit that these mediations have real effects in the legal sphere.18
The term “mule” as synonymous with a drug courier is also not without contention, particularly because of the racial and class-based implications it carries.19 For instance, in an extensive breakdown of potential roles within drug trafficking networks, a transnational drug treatment organization categorizes mules as the agents with the lowest degree of organization (dependent on others for orders) and the lowest investment in the trade (focused on the service rather than the ultimate selling of drugs).20 These connotations are partly based in the lived realities of those who smuggle drugs across borders. Acting as a drug mule is a profitable endeavor for under- or unemployed workers in developing countries. In many places, it functions as a gig economy where people take these jobs sporadically as supplemental income.21 Likewise, while the targeting of women of color for inspection results from implicit or explicit racial biases, captured mules sometimes learn how to mobilize racial prejudices to their advantage.22 The narratives about who counts as a drug mule and why are not merely problematic stereotypes; they are crucial loci of symbolic and material struggle over how bodies are marked, sorted, and disciplined.
Located at the intersection of these gender, class, and race discourses and characterized by her role as mere conduit within a broader network of transnational flows, the drug mule acts a crucial node of mediation from which to examine the assemblage of narcotrafficking. She stands at the intersection of various strands of analysis precipitated by the phenomenon, including the precarity of bodies, the instability of borders, and the influence of popular media portrayals.23 Despite being specific to narcotrafficking, this figure also proves instructive to trace contemporary shifts in subjects’ interactions with digital technologies at a transnational scale. The drug mule functions as an aperture into the broader issues precipitated by the datafication of borders and bodies because, ontologically, this figure only exists when bodies serve as conduits for contravening enforced borders. The drug mule's travels render, and undo, what can be termed the datalogical border-body entanglement of our contemporary moment.
THE DATALOGICAL BORDER-BODY ENTANGLEMENT
Current developments in border studies replace the once static perception of the national boundary with a highly differentiated entity, refiguring the border as something other than a fixed, stable line. Borders are constructed by negotiations between individuals and institutions, performed by habitus, and subject to instability and constant change. Because of this, borders become one of the sites where the turbulence and conflictual intensity of global capitalist dynamics are particularly apparent.24 In symbolic and material terms, borders function as complex social institutions, constituted by tensions between practices of border crossing and border reinforcement.25 Saskia Sassen argues for globalization as the force driving “the actual and heuristic disaggregating” of the border when, on the one hand, global economic forces creep into national territories while, on the other hand, national sovereignty apparatuses extend to places around the globe.26
This actual and heuristic disaggregation is increasingly enabled by information and communication technologies. ICTs help state agents fortify national boundaries from afar, generating network-like borders that “jump scale” into transnational space and “touch down” in various nodes across the globe.27 These network-like borders depend on complex databases and biometric technologies located in distant territories that identify, profile, and track inbound travelers. Stephen D. N. Graham emphasizes that the role ICTs play in border sorting and surveillance is “to add friction, barriers, or logistical costs to the mobility and everyday lives of those deemed by dominant states or service providers to be risky, unprofitable, or undeserving of mobility.”28 In fact, because of ICTs, not only are goods, information, and people flowing across the world, but borders themselves are on the move. Beyond their physical manifestations in checkpoints, fences, or walls, borders rise wherever “concretions of power struggles in a specific space [are] materialized within a territory.”29 Conceptualizing the border as a transnationally interconnected entity suggests both an expansion of the locales considered part of the border and an extension of the technics counted as practices of border making. The paradox at the heart of contemporary bordering practices is the simultaneous, yet sometimes contradictory, tendencies of border concretion and border disaggregation.
What these arguments illuminate is the processual nature of the ICT-enabled border. The increasing dependence of nation-states on information technologies for the maintenance of their geopolitical divisions grants borders a newfound digital ontology. “Where the digital and algorithmic is concerned,” argues Olga Goriunova, “the ontological question of ‘what it is’ becomes the question of ‘how it works.’”30 In order to understand contemporary borders it is crucial to trace the algorithmic processes that organize and shape these digital networks. How a border is constructed also depends on a meshwork of policies, practices, technologies, infrastructures, and a variety of human and nonhuman agents. Indeed, the success or failure of narcotrafficking enterprises depends on smugglers’ knowledge of border enforcement practices, on drug mules’ attempts to circumvent such practices, and on state agents’ efforts to outwit these circumvention attempts. This data-driven border-making assemblage is not only performative but also enactive. The ICT-enabled border comes to be only when, and because, certain flows pass (or not) through it.
The disaggregation of the border is, therefore, tied to the datalogical becoming of the body. A physical blockage prevents a body from crossing the geopolitical boundary only in the last instance. Borders can “jump scale” because data-driven body sorting mechanisms enable or prevent movement long before this physical encounter. Like the contemporary border, in the current phase of technogenesis the body becomes what Mark B. N. Hansen terms a body-in-code, one “submitted to and constituted by an unavoidable and empowering technical deterritorialization.”31 While its materiality remains central, within this datalogical turn, as Patricia Clough et al. term it, the definition of the body must include “bodily practices [which] themselves instantiate as data, which in turn produces a surplus of bodily practices.”32 A body is not simply a physical organism, but an assemblage of physical and virtual characteristics that become constituted and deconstructed at particular instances. In the case of the drug mule, someone like Rosa de Montoya Hernandez would not arrive at the Los Angeles airport without being screened when buying a flight ticket and boarding the plane.
Smugglers must now contend with ICT-enabled border checkpoints long before arriving at the geopolitical boundary. Similarly, since drug smuggling by swallowing requires multiple tactical uses of the mule's body, the datalogical becoming of the body informs the ways these tactics are enacted. Mules must not only have no previous criminal offenses, but indeed a spotless data record in order to avoid detection by the surveillance systems used by the state's security agencies. Thus, this assemblage is not only a technological or media-based one, but also a “racializing assemblage,” what Alexander Weheliye terms the technical and institutional norms that construct and maintain differences between individuals by inscribing these differences onto physical bodies.33 Despite the developments in technical means, institutional logics continue to render drug mules as specific types of individuals, marked along class, race, and gender lines. The protocols derived from these logics carry within them historically constituted assumptions that, in turn, continue to perpetuate these biases.
The datalogical becoming of the body, therefore, is not value-neutral. Data-driven classification systems also become sites of political work. Since systems that purport to recognize biological features function through standardization, the data captured and normalized in facial recognition analysis, for example, then “gets pushed back onto faces, as the standardized coding system comes to define what facial expressions mean and how they work.”34 The datalogical becoming of the body results not only in a proliferation of virtual information of an individual but also in a reinscription of informational codes onto the physical manifestations of individuals. Security systems may flag travelers whose economic means do not match their planned trips or those with no previous history of transnational travel. Knowing these protocols allows smugglers to plan against them. In fact, most successful mule recruiters possess a sophisticated notion of intersectionality, hiring women who fit into a “more normal than normal” schema, with specific ages, skin colors, ethnicities, and experiences of traveling that are unlikely to flag any automated system's filters.35
Finally, for the swallower, the physical body itself with its inherent biological affordances dictates the temporalities and affects—and success—of their border crossings. The contemporary mediation of the body is thus characterized by its fragmentation within the informational realm and its subsequent reinscription as material markers of difference. Embodiment results from the convergence of material features, social practices, and symbolic charges attached to it. Amid this border-body datalogical entanglement, the mule's drug-carrying body, in its computational and biological forms, has to move through a series of physical and virtual checkpoints as it travels through modern transnational border networks.
How this travel occurs must also be conceptualized. Concretions of power materialize both at the points where borders are enacted and at those where bodies are constituted. If “big data is moving data,” then it “cannot be captured or held static or it would lose its very value,” socially, economically, and politically.36 The data that constitutes borders and bodies is always in flux—that is, systems are continuously processing data that allows or prevents the movement of bodies across previously programmed boundaries. At the same time, this data is only meaningful because it is eventually reinscribed onto physical entities. Processing the virtual bodies of travelers matters because it determines whether physical bodies can move across borders. CBP protocols, such as automated systems filtering and agent inspections, produce borders at the same time that they inspect virtual or physical bodies. It is the co-constitutive nature of these complementary practices that further crystallizes the datalogical border-body entanglement.
Herein lies the historical distinction of the contemporary form of border making and body marking. In an ICT-driven world, the materialities of the border and the body become effects of the dynamics of power inherent in the pervasive use of data processing mechanisms. As Judith Butler succinctly puts it, “The matter of bodies [and borders] will be indissociable from the regulatory norms that govern their materialization and the signification of those material effects.”37 Digital technologies of data capture codify which bodies warrant inspection beyond the usual checkpoints for border crossing. Demographics and a subject's digital trail increasingly feature as markers that signify a body's ability to circulate. The filters for these forms of sorting are indissociable from regulatory norms, particularly those norms enacted by the state as a way to reassert sovereignty. In the border-body datalogical entanglement, there is no making of borders apart from the reading of bodies—a feature that both reaffirms the methods of capture enabled by ICTs and points to potential avenues for their tactical undoing.
CBP PROTOCOLS AS BORDER-BODY ALGORITHMS
Data about the border or the body does not preexist the means of its capture. Whether digital or analog, the procedures that process data contribute to data's emergence and constitution. As Johanna Drucker argues, “Data are capta, taken not given, constructed as an interpretation of the phenomenal world, not inherent in it.”38 Drucker's observations here speak to the use of computational tools to organize and visualize data in forms comprehensible to human audiences. The latter part of this article, however, tests this insight against the case where the human functions as the tool that organizes and makes comprehensible capta from the phenomenal world. In particular, the CBP officer who follows a strict logic-driven, step-by-step procedure—an algorithm—works as the primary processor of data about the body at the border. The shift toward understanding logical procedures processed by humans as akin to the datalogical entanglement driven by ICTs is not merely a metaphorical exercise. Instead, it is an attempt to challenge the exceptionality of digital border securitization by foregrounding its analog continuities. Although the border-body datalogical entanglement provides a new platform for obdurate biases about which bodies are marked as criminal at the border, it is also crucial to reveal how these logics remain active in non-computational forms alongside newfound digital modes of capture.
A datalogical drug mule is thus constituted through the performance of both computational and non-computational algorithms followed by US Customs and Border Protection. Importantly, this performance is bound to remain imperfect because, as Adrian Mackenzie and Theo Vurdubakis argue, the execution of an algorithm is “a fraught event,” always “mired in ambiguity, undecidability, and incompleteness.”39 The noncoincidence between the symbolic structure of a code and its enactment, the slippage between knowing and doing, resides in the performance inherent in executing said code. Mackenzie and Vurdubakis argue that the iterability required by performativity is itself coded. The constitution of a code requires constant enactment. That iterability and recursivity are constitutive of algorithmic execution, that “code works by being coded,” means that performance and excess remain intrinsic to the application of algorithms and, therefore, to the meaningful processing of data.
When considered as non-computational algorithms, CBP protocols illustrate this link between performance and data capture. In the case of drug mules, performance becomes a crucial aspect to their success, since they actively perform to avoid detection at two key points in their air travel. First is the pre-departure TSA screening. Here the CBP depends on a version of interpersonal deception theory (IDT), a method that purportedly uncovers what airline passengers are hiding by reading their body signals. Initially, this technique was controversial for the possibilities it opened up for implicit racial and gender biases. Even as these concerns were glossed over, though not fully addressed, IDT has never achieved the unqualified success desired by the Transportation Security Administration (TSA). Despite attempting to develop “a universal system for decoding the signs or nonverbal tells of deception,” experts in IDT never “altogether escaped semiotics,” therefore allowing “for skilled performers to train for and rehearse unwitting performances of transparency.”40 Passengers might submit to what Rachel Hall terms “performing transparency” in order to have a more expedient passage through the various airport security mechanisms. That is, travelers may acquiesce to a “voluntary transparency, or demonstrate readiness-for-inspection,” by “‘clearing’ their opaque bodies, bags, and belongings.”41 People attempting to avoid detection, such as the drug mule, can also participate in this performance, thereby mobilizing the semiotic failures of the security apparatus for their own aims. In addition to the pre-departure TSA screening, the drug mule must also perform transparency at the migration and customs checkpoints upon their arrival. It is at these latter points that the mule most forcibly contends with the guidelines that US Customs and Border Protection has for finding and detaining smugglers.
The CBP's list of criteria for detaining a passenger consists of forty-five markers that a border patrol agent must look for in a passenger's appearance and comportment.42 While some correspond to bureaucratic measures, such as lacking the proper permit or an expired passport, the bulk of these criteria have to do with physical attributes or behavioral cues. Many of these selections are confusing, if not contradictory. The handbook suggests detaining a traveler if they are “overly talkative and friendly” or if they are quiet with “a poker face.” Likewise, exhibiting nervousness is as much a red flag as being “unusually cool, polite, or cooperative.” The assessment of what a poker face is, or at what point one becomes “overly friendly” or “unusually cool,” remains unclear. The expectations for how travelers should act during this process are therefore impossible to meet. Further, these security measures assume a certain predisposition of travelers toward authority figures. For populations with historically inflected “negative and antagonistic relations with law enforcement agents,” their wariness, fear, or defensiveness may often have nothing to do with the specific instance of meeting the customs agent, but with a general disaffected orientation toward policing structures.43 The performance of this security code thus allows for a wide range of contingencies to emerge in its continuous executions.
Key among these contingencies is the seeping of racial, class, and gender biases into the performance of the border patrol agents. In the case of IDT, accusations of profiling have plagued the program since its trial run at Boston's Logan Airport.44 The purported impartiality of the system for detecting deception falls apart in practice, when the relationality between the watcher and the watched exceeds the normative expectations of the code. Similarly, the CBP algorithm for detaining passengers is plagued with latent codifications of what a normative passenger should be and who is already a questionable body in movement. For instance, many of the qualifications mark a passenger if they have “taken a very expensive trip not commensurate with economic station in life” or is “overdressed for company.” As well, if a traveler has “had beard, moustache, or sideburns recently shaved off” or is “out of character,” then the handbook codes these as red flags. Class-based assumptions about who is an appropriate traveler, and in what ways they must behave, are already introduced in the code of the security algorithm. The almost contradictory directions within this handbook place the enactment of the instructions entirely at the discretion of the border patrol agents.
At the same time, the excesses emergent from the algorithm and its performance can also become a tactic for those trying to bypass the security apparatus. The CBP handbook guidelines for detaining passengers are publicly available through Freedom of Information Act (FOIA) disclosures, though there might be ongoing changes to the latest version. It is not unconceivable that recruiters looking for possible mule recruits would take into consideration the recruit's ability to fool the code for detecting suspect passengers, and perhaps train them to perform transparency adequately. In fact, there is evidence that some already do. Criminologist Jennifer Fleetwood's interviews with drug mule recruiters in Ecuador reveals that they take into account the assumptions inherent in the CBP's guidelines when finding possible smugglers, even if there is no evidence that recruiters have ever seen this handbook. Recruiters also take into account assumptions that security agencies may have about the gender, age, and race of the smuggler, preferring to draft men, seniors, and European-looking Latinos—a distinct contrast to mainstream ideas about a drug mule as young woman of color.45 Notably, many recruiters approached people who were already employed, thwarting the classed assumptions in the CBP handbook that a mule would be a blue-collar worker pretending to afford a high-cost trip. This characteristic proves more tactically advantageous given the datalogical becoming of the body. If airline passengers are entered into the system that screens them before they arrive at their destination, choosing people who are employed lends credence to the image that the smuggler is only a person traveling for leisure.
In this regard, recruiters’ and smugglers’ responses to the security apparatus they are set up against reflects the broader experience of user adaptation to algorithmic interactions. As algorithms increasingly permeate aspects of everyday life, users not only take them up but also rearticulate them. By affecting how people seek information and perceive the contours of knowledge, algorithms shape how users understand themselves. Understanding themselves differently because of this algorithmic mediation, digital users change their practices in turn. The interaction between algorithmic performance and user adaptation functions as a recursive loop, where the calculations of the algorithm respond to the tactics of the users, and vice versa. Further, as algorithms change to respond to their users, the users’ tactics also evolve and adapt. People can make themselves “already algorithmically recognizable,” orienting themselves to the dynamic, complex processes that sort them and their actions and making decisions informed by knowledge of these processes.46 Insight into the workings of algorithms, understood both as static code and as dynamic performance, becomes a modern form of power struggle.
Understanding performative relations as well as procedural logics allows for attending to power and technics in tandem. If people, algorithms, and institutions always act in relation to each other, engaging in complex, dynamic negotiations, then attending to how algorithms are enacted reveals far more than a technical inspection of the underlying codes. An analysis of the algorithmic struggle between people, institutions, and security apparatuses must begin with a consideration of the affordances, iterations, and contestations latent in the algorithmic entanglement. The agon of algorithms, the fact that interacting with computational and non-computational protocols implies the continued potential for struggle, is a feature of contemporary techno-mediated society, one that becomes particularly salient in regard to surveillance and securitization algorithms.47 In the case of drug mules, this agonistic model of algorithms applies not only to the code of detection but also to the procedural logic of detention. That is, the power relations between smugglers and state agents occur not only in the performance of transparency at a pre-departure TSA screening or at a customs inspection post-arrival. Instead, they continue to be articulated even if drug mules are suspected and detained by border patrol agents.
THE BIOLOGICAL STRIKES BACK
The CBP handbook's procedure for detaining and inspecting passengers suspected of carrying drugs within their bodies operates in step-by-step order.48 Two key aspects of this non-computational algorithm are noteworthy. First, the procedure is timed. The clock begins from the moment the person is asked to undergo a pat down by a CBP official. Then, every two hours, CBP officials must report to their supervisors to request continued detention if no wrongdoing has yet been proven. After eight hours, Immigration and Customs Enforcement (ICE) must contact the US Attorney's Office to consult whether it is advisable to continue detaining the passenger and to pursue the legal requests to do so. The second noteworthy feature is the ongoing feedback loop within the algorithm. CBP officials shepherd the suspected passenger through the process, but these officials must continually check with supervisors, legal aides, and medical experts as decisions are made over whether and how to proceed. Finally, the workings of the procedure depend on the detained person's consent. Whether the suspected smuggler consents or not determines the subsequent steps taken, the outside resources called for, and the timing of the process itself.
In fact, it is at the moment of capture where the agonistic dimension of performing algorithms is most salient. While CBP officers follow the step-by-step instructions for detention and inspection—there are, in fact, ten regular numbered steps and two final ones (release or arrest)—the algorithm's mechanics are not determining of the process's outcome. Instead, it is the performance of this algorithm and the resistance or acquiescence of the agents involved (officers, supervisors, medical professionals, and the detainees themselves) that shape the possible and eventual outcomes. In the case of the suspected smuggler, for example, the denial of consent at every other step determines one of the many potential routes the procedure can follow. The CBP does not require the passenger's consent for a pat down. Likewise for a partial body search, where some items of clothing are removed and the passenger can decide to take the clothes off voluntarily or not. A woman detained by CBP can deny consent to a pregnancy test (prior to an X-ray), an X-ray, or a body cavity search.49
The X-ray screening is another instance where the datalogical meets the biological. Paradoxically, although the X-rayed body, “stripped of its overinscribed gender- and race-encoded epidermis and organs,” may be the sole point in the border-body entanglement where markers of difference are irrelevant, it is also a moment when the data-body encounters a new surveillant technology.50 The question of mediation remains. Radiologists, for instance, struggle with questions of how clearly drug pellets can be detected in an X-ray. Plain abdominal radiography, the most commonly used method, may not be sensitive enough for carriers with a small number of packets. False positives occur in this method because of bladder stones, inspissated stool, or intra-abdominal calcifications. In fact, radiologists focus not on finding the drug pellets, but on the anomalies that signal their presence such as a “rosette-like finding” formed by air trapped in the knot where a condom is tied or a “double-condom” sign formed by air trapped between layers of latex.51 The data provided by an X-ray are not self-evident; they must be continuously interpreted by medical experts.
Although the smuggler can deny consent for an X-ray, the port director can overrule the denial of consent for a body cavity search. Historically, these rulings have proven ineffective. In the late 1980s, surveys found that 80 to 85 percent of women subjected to body cavity searches were found to carry no drugs. This overreliance on an intrusive method with an embarrassingly low success rate is what some constitutional law scholars diagnose as a “drug war exception” to the Fourth Amendment.52 Under the increasingly militarized state, the mandate to protect the geopolitical border supersedes the legal and ethical need to respect the body's boundaries.
As an alternative to a body cavity search, and in consultation with general counsel, the port director can instead approve a detention for monitored bowel movements (MBM). In this case, the passenger is taken to a medical facility to be observed for a period of time—at least until the eight-hour mark—to see how she reacts to a prolonged detention and whether her metabolic functions produce the evidence needed to convict. The MBM stage is the last step in the algorithm as detailed in the CBP handbook. It either provides the necessary proof to convict, or it does not and the passenger is released. As a step in the process, the MBM remains a black box, but as a performance of the algorithm, it marks the last instance for agonistic power relations to occur. It signals the reemergence of the physical body within the border-body datalogical entanglement, particularly the contest between the physical body's temporality and the temporality dictated by the algorithm itself.
The eight-hour mark serves as a crucial turning point when CBP officials and the US Attorney's Office must decide to prolong detention or bring charges against the suspected passenger. If there have been no bowel movements by this time, there is little evidence to support an arrest, but minor affective signals can provide incentive for prolonged detention. The conflict in the smuggler's body lies in controlling biology longer than the security procedures take to unfold. Numerous variables are at play: individual metabolisms have their own rhythm, and some people can train in advance to slow theirs down; stress can accelerate bowel movements, but continued resistance to defecation, on the other hand, signals the body to temporarily shut down intestine functions. In addition to the case of Rosa Montoya de Hernandez, there is that of Delaney Abi Odofin, who—according to legend—spent days in detention before passing anything at all.53 When individual bodies’ resistances compete with legally mandated temporalities of detention, the struggle between these contesting forces produces multiple potentially unpredictable outcomes.
The return of the biological body into the algorithmic processing of suspected drug mules provides another stratum to the border-body datalogical entanglement. Mules can resist emitting incriminating signals, but their bodies may betray this data beyond the mule's control. Likewise, CBP agents have no standard method for analyzing “superficial bodily responses” and must interpret body signals within each specific context of detention. This inescapable biological dimension of the datalogical drug mule echoes recent scholarship arguing for including critical analysis of biology in feminist critique.54 In particular, it speaks to the importance of considering metabolism, and the gut, alongside questions of social difference.55 Further research into how individual physiology affects the emission of actionable data from a suspected smuggler's body could illuminate this biological aspect of the border-body datalogical entanglement.
The physical body and its metabolic functions become the last border through which drugs must pass for trafficking to be successful. The border-body datalogical entanglement is, therefore, also a physical and biological entanglement since materiality persists despite datalogical dispersions. This final instance illustrates how the potential for countering algorithms becomes further marked in embodied performances. It is the drug mule's ability to resist passing any pellets before being released by CBP that literalizes the concretions of power emerging at border points, the agon of being already algorithmically recognizable, and the importance of focusing on mediating processes to understand how narcotrafficking becomes differentially embodied.
Ultimately, this investigation into the corporealization of the datalogical drug mule illustrates a number of features about the ICT-driven world of border administration and bodily regulation. First, it reveals the border-body datalogical entanglement as a new site of struggle over the definition of geopolitical and embodied boundaries. At the same time, it alludes to how analog practices continue to influence which bodies are marked and allowed to circulate. The case of drug mules foregrounds the multiple ways in which subjects learn to become algorithmically recognizable, engaging with and undermining algorithmic forms of capture by the securitized state. Finally, this analysis emphasizes the matter of bodies: how the biological dimension of the human body continues to play an indispensable, and irreducibly material, part in the agonistic performance of data capture. In her travels across computational and non-computational networks, through digital and biological checkpoints, the datalogical drug mule reaffirms and undermines the data-driven security apparatus of the state, leaving behind a trail of openings for its potential undoing.