A. Introduction
While long taken for granted, the presence of gender markers on identification documents, such as passports, has faced increasing scrutiny in recent years. Scholars and activists have highlighted the negative impact these markers can have on gender non-conforming individuals and questioned the efficacy of using gender information as a tool for verifying identity.Footnote 1 Additionally, some have demonstrated that people’s experiences of harassment and discrimination triggered by gender markers are also shaped by how they are positioned in other systems of inequality, such as race and religion.Footnote 2 With the aim to follow up on debates that have recognized the unequal effects of gender markers on passports, I discuss how they function as a tool of current migration control in the European Union (EU). I suggest that gender markers on passports are implicated in upholding the digital racial border by helping to single out “unwanted” migrants.
Although much has been and could be written about the EU’s structurally unequal and discriminatory migration control, especially given the evolving algorithmic border governance, this Article focuses on one specific element: gender markers on passports. This specific angle aims to show that gender markers on passports are not only by-products of normalized cisnormativity but also an instrument to control the mobility of people based on racialized lines. In particular, I analyze how gender markers on passports are used as input factors in digitalized risk assessments in migration governance that help produce a racially desirable biopolitical body. However, this specific angle also comes with limitations. For instance, my focus on the intersection of gender and race in the construction of “the unwanted” migrant does not do justice to the various ways class and other systems of inequality, such as caste, play a role in mobility. Moreover, while the Article brings together literature on gender, migration, race, and technology, the challenge of this cross-disciplinary exercise is that certain complex debates can only be addressed briefly.
The Article focuses on the role of passports in border governance in the region of my own origin, the EU, due to ongoing efforts to further digitalize the borders of “Fortress Europe.”Footnote 3 While it highlights how gender markers on passports contribute to the reproduction of racial hierarchies in the EU, it also demonstrates that it would be overly simplistic to assume abolishing such markers would necessarily combat border racism and render passports gender-neutral. While abolishing gender markers on passports could remove one data point used in the policing of bodies—and might therefore represent a valuable reform—the adoption of alternative identification tools, such as biometric technologies, could reproduce similarly racialized and gendered borders. With the abolition of gender markers on passports, the gender binary would not lose its grip on the racialized border and its technologies, given their deeply racialized and gendered nature. My main message is, therefore, that discussions on gender markers in passports must consider their productive effects on not only gender relations but also global race relations, particularly through their role in migration control.
Before delving into my analysis, let me say a word on terminology. I use the terms “sex” and “gender” interchangeably, because relevant documents and policymakers often do the same in discussions on identity cards.Footnote 4 As will become clear, in the context of identity documents, gender and sex are both legally constructed categories that are recorded for specific purposes, such as migration governance, rather than pre-legal descriptors of a natural state of bodies.
The Article continues by explaining the concept of digital racial borders and how they are being enshrined into the EU’s migration governance. It then examines the emergence of gender markers on EU passports as part of efforts to digitalize border controls for security purposes. This is followed by an analysis of how gender markers on passports play a role in three border governance tools employed by the EU: Advanced Passenger Information (API), Passenger Name Records (PNR), and the Visa Information System (VIS). Before concluding, I will reflect on what passports without gender markers would mean in the future of biometric borders.
B. Digital Racial Borders in the European Union
With the rise of information technology, borders are becoming increasingly digital rather than merely physical. In addition to turning people away at the physical border, sometimes through extreme physical violence such as pushbacks, nation states are becoming guarded by many layers of online borders. The legal identities of migrants are thereby regulated and surveilled through secretive algorithms, big data, biometric technology, and drones, in addition to armed security forces, which makes Martina Tazzioli argue that “migrants are haunted by an imbrication of digital and non-digital traces and documents.”Footnote 5
The progressive reliance on technologies in border management has been described through several concepts, including “virtual borders,”Footnote 6 “biometric borders,”Footnote 7 “virtual fence,”Footnote 8 and “datafication of migration.”Footnote 9 The centrality of technology to determine “the suspicious” has significantly accelerated since the war on terror.Footnote 10 As Louise Moore explains, in the aftermath of 9/11, the United States (US) focused on developing risk profiling techniques via information technology to classify people into degrees of riskiness based on different components, such as student status, gender, and religion.Footnote 11 While initially driven by the US, the enhanced use of digital tools to govern borders and assess the riskiness of people has arrived in all other regions, including the EU,Footnote 12 constituting the basis for Henk Van Houtum’s claim that “the EU’s external border regime is transforming into a digital and selective border machine.”Footnote 13
Given the increasing digitalization of borders, E. Tendayi Achiume has developed the concept of “digital racial borders.”Footnote 14 This extends her earlier work on racial borders, which, alongside other scholarship,Footnote 15 such as Radhika Viyas Mongia’s research on the racist origins of passports,Footnote 16 demonstrates that current migration control is deeply racialized.Footnote 17 Achiume traces the origins of current mobility restrictions to the colonial era, when movement was free for Europeans but tightly controlled for enslaved and colonized people. With decolonization and the recognition of sovereignty as a principle governing the relations between states, such restrictions could no longer be justified overtly on the basis of race. Borders were established and it became the prerogative of nation states to protect these borders based on “race-neutral” criteria, such as nationality and passports.Footnote 18 Yet, rather than serving as neutral criterion, Achiume argues that “nationality within immigration regimes remains a mostly bulletproof mechanism for racialized exclusion and differentiation.”Footnote 19 Accordingly, today’s borders are a continuation of colonial structures by conferring the privilege of mobility, such as visa-free travel, to those most proximate to Whiteness.Footnote 20 Nationality has thereby become the purportedly race-neutral criterion on the basis of which racialized hierarchies are justified.
The EU’s external borders illustrate how borders are increasingly digitalized and are often cited as examples of racial borders. Critical migration scholars have, in particular, scrutinized the EU’s visa regulations, which divide third countries into a “negative list” and a “positive list.” Individuals with passports of countries on the negative list require visas for entering the EU, whereas nationals of countries on the positive list are exempted from visa requirements for stays of less than three months in any 180-day period.Footnote 21 These lists were once officially—and are still often informally—called the “Black” and “White” lists, but their formal names have been changed due to concerns about racist connotations.Footnote 22
As several scholars have observed, it is striking that the two lists follow clear racial lines.Footnote 23 Countries with predominantly Black or Brown populations, such as all African states except for Mauritius and Seychelles, and South Asian states are largely on the negative list. In contrast, the positive list is dominated by countries in North America, Eastern Europe,Footnote 24 Oceania,Footnote 25 and South America.Footnote 26 About 80% of countries on the positive list have a predominantly Christian population, while only a few countries with Muslim majorities are exempted from visa requirements.Footnote 27 Israel and the United Arab Emirates are the only Middle Eastern states on the positive list. Given the overrepresentation of countries whose nationals are predominantly White, White-passing, and/or Christian on the positive list, EU visa regulations are in effect racialized despite being formally race-neutral.
The digitalization of the EU’s external borders has been accelerated since, at least, the EU Commission’s communication on “Smart Borders” in 2011. This communication highlights several instruments that could be strengthened to support “smart borders,” including API, PNR, and the VIS.Footnote 28 In 2019, the interoperability framework was established with the aim to connect six large-scale IT systems recording data on third-country nationals under the EU Agency for the Operational Management of Large-Scale IT Systems in the Area of Freedom, Security and Justice (eu-LISA).Footnote 29 The databases include the VIS, the Schengen Information System (SIS), Eurodac,Footnote 30 the European Criminal Records Information System for Third-Country Nationals (ECRIS-TCN), as well as the newly established Entry/Exit System (EES) and European Travel Information and Authorisation System (ETIAS).
A key driver behind the interoperability framework took place in the summer of 2015, when the media, the public, and politicians talked of a “refugee crisis” in Europe. While pockets of solidarity with those fleeing their home countries initially emerged across Europe, the time was also marked by a rise of anti-migration discourse, which found wide support when terrorist attacks occurred in Paris in late 2015.Footnote 31 The interoperability framework became concretized as part of the attacks’ direct aftermath. It reinforced the linking of migration governance with security concerns in the EU, such as by connecting databases on migrants’ legal status in terms of visas (VIS), travel authorization (ETIAS), and asylum claims (Eurodac) with those focused on crime prevention (ECRIS-TCN and SIS).Footnote 32 Efforts to further digitalize the EU’s external border through expansion of API, PNR, and the VIS must thus be also understood as part of broader developments to strengthen the migration-security nexus in the EU.
C. Gender Markers on EU Passports as Digital Advancements in Border Control
Gender markers have not always existed on passports. In fact, many states across the world, including EU states, did not display gender in the form of a boxed category on their passports until the last quarter of the twentieth century. For example, Belgium passports used in 1980 did not show gender markers and Italy introduced gender markers on passports at some point between 1975–1985.Footnote 33 The systematic inclusion of gender markers on passports, which began after 1980, was significantly driven by digital developments of that period. Advances in computer science made it possible for machines to read passports, which led to the creation of machine–readable passports (MRPs), and with them the introduction of gender markers on passports. Gender markers thus emerged on passports often together with the digitalization of identity controls at borders and elsewhere.
The work of the International Civil Aviation Organization (ICAO) significantly influenced the adoption of MRPs displaying gender markers since the 1980s. The ICAO published its standards on Machine-Readable Travel Documents (Doc. 9303) for the first time in 1980. While the first version of Doc. 9303 recommended states to note sex as one of the binary characters—F or M—on passports,Footnote 34 the ICAO introduced the additional option to display sex as an X standing for “unspecified” in 1996.Footnote 35 By now, most states follow Doc. 9303 and mention gender in the form of either an F or M, or, in exceptional cases, an X on their MRPs. The few EU member states that have recognized gender categories other than female or male, such as Germany and Austria, display these categories as an X on passports.Footnote 36
Prior to the publication of Doc. 9303, international passport regulations did not recommend the display of gender markers on passports. For instance, the first uniform type of passport adopted by the League of Nations in 1920 did not display sex markers for the main passport holder, but sex markers were included only for children traveling on the passport.Footnote 37 The shift regarding gender markers thus occurred when states started to employ machines to read passports, which made it necessary to encode personal information on passports into standardized characters. While qualitative information that was difficult to encode, such as eye color and special characteristics, were included in the 1920 uniform passport model, they disappeared from the MRPs proposed by Doc. 9303.Footnote 38 Instead, sex must have been perceived as a characteristic that was universally understood as a binary and could therefore be easily turned into the machine-readable characters of F and M. Drawing on narrowly boxed identity categories, such as sex, for identification at borders, rather than more complex descriptors of physical appearance, was thus also the result of technical developments.
The adoption of MRPs, which also intensified the use of gender markers as elements of identification, was undertaken with the aim to make border controls more efficient and secure. It occurred in the context of emerging anxiety over migratory flows from the Global South to Europe in the 1980s.Footnote 39 When the ICAO recommended that states establish MRPs for the first time in 1980, it did so with the aim to speed up border controls, given the overall increase of travel in the period of economic recovery after the war.Footnote 40 Yet, Martin Lloyd argues that it is debatable whether MRPs truly made border checks faster, but they undoubtedly facilitated surveillance by allowing border officials to cross-check a traveler’s identity within seconds against databases held by various state bodies, including the police, intelligence services, and tax authorities.Footnote 41 Indeed, while not explicitly stated in the first version of Doc. 9303, the ICAO nowadays describes MRPs as enabling “improvements in security, clearance speed and accuracy of verification.”Footnote 42 This underscores that the inclusion of gender markers on MRPs was also tied to efforts to improve border efficiency and security through digital technologies.
Enhancing security was partly the rationale for requiring EU member states to issue ICAO-compliant passports. After discussing the idea to establish a passport union for a few years, EU member states adopted a resolution on a uniform passport model in 1981.Footnote 43 In this resolution, it was agreed that the European uniform passport will mention sex and comply with ICAO recommendations.Footnote 44 This non-binding agreement was followed by a binding regulation on biometric passports in 2004, making it obligatory for EU states to issue MRPs in accordance with Doc. 9303 and thus with gender markers.Footnote 45 While the 1980 Resolution was adopted with the aim to facilitate travel and strengthen a feeling of community among EU nationals,Footnote 46 the 2004 Biometric Passport Regulation stresses the security benefits of issuing ICAO-compliant MRPs.Footnote 47 Thus, gender markers on EU passports gained traction with the rise in the use of digital tools to check travel documents for the purpose of enhancing border security.
D. Using Gender Markers in Risk Assessments to Protect Digital Racial Borders
Having shown that gender markers on EU passports emerged with the digitalization of passport controls in the 1980s, I now analyze their current role in protecting the EU’s racial borders through risk assessments. I first noted the use of gender markers in border-related risk assessments in a review of gender markers on passports presented by New Zealand to the ICAO in 2012. This review supports the retaining of gender markers on passports by arguing that “[k]nowing a passenger’s gender helps improve security as it allows border authorities to categorise passengers and do risk assessments before they arrive”.Footnote 48 This reference indicates that states have long used gender markers on passports to identify “risky” bodies.
My subsequent analysis of API systems, PNR data collection, and the VIS demonstrates that gender markers on passports are increasingly employed in risk assessments conducted at the EU border. This renders it more improbable, albeit not impossible, that they will soon be eliminated. In the near future, gender markers on passports will therefore likely continue to police gender presentations in line with cisnormative ideals, often resulting in discrimination and scrutiny of gender non-conforming persons.Footnote 49
My analysis further shows that gender markers on passports do not only affect gender non-conforming persons, but are also used to identify other gendered groups deemed “risky” or “unwanted.” Since the war on terror, gender and race often intersect in the construction of those classified as dangerous, having in particular negative effects on men of color, especially men connected to Muslim-majority states.Footnote 50 The former UN Special Rapporteur on Human Rights and Counter-Terrorism, Martin Scheinin, documented risk profiling operations that relied on proxies for race, such as nationality, in connection with other statuses like gender. For instance, he deplored a US operation targeting men of certain ages and nationalities as well as the German Rasterfandung, which classified all Muslim men aged 18–40 years who were current or former students and born in or nationals of Muslim-majority countries as potential terrorist sleepers.Footnote 51 These operations undertaken in the name of security provide evidence for Hall and Clapton’s claim that “risk is not a neutral or natural ‘thing’; rather it is fundamentally embodied, racialised, and gendered.”Footnote 52 As demonstrated below, gender markers on passports assist in the construction of risk as gendered and racialized.
Yet, one clarification is in order. While I will show that gender and race become meaningful in relation to each other in border-related risk assessments, I do not suggest that they carry the same weight in the construction of the racialized “risky” migrant, nor that such assessments have identical effects across different groups, such as Muslim men and gender-nonconforming individuals. Rather, my analysis is limited to demonstrating that gender markers do play a role in this construction, while more qualitative research and insights into highly secretive algorithms are needed to show how they matter concretely.
I. Constructing Suspicious Bodies Through Advanced Passenger Information
A nowadays widely used border security measure that often relies on gender markers is API systems.Footnote 53 API systems require airline companies, and sometimes shipping companies, to share biographical data and specific journey-related information about passengers and crew members with border authorities before an aircraft or vessel’s departure and arrival. States’ primary aim in receiving API is to check the arriving passengers against terrorist watchlists and other security databases.Footnote 54 API systems emerged largely after 9/11, with the US making API-sharing mandatory almost immediately after the terrorist attacks and the EU adopting a directive requiring airline companies to share API data, upon request, in 2004 (2004 API Directive).Footnote 55 By now, both the UN Security Council and the ICAO request states adopt API systems.Footnote 56
Two new EU regulations adopted in 2024 require SchengenFootnote 57 member states to collect sex as part of API data (API Regulations).Footnote 58 One regulation focuses on API data collection for facilitating border checks by combating “illegal immigration” and security threats, while the other one regulates the use of such data for addressing terrorist offenses and other serious crimes. This means that, in the EU, sex information is collected as part of API data to detect both people who are suspected of conducting crimes and those who might become undocumented migrants.
While the 2004 API Directive required states to establish some form of API system through which airlines could be requested to send API data, the Directive defined only a minimum set of data to be transferred, which notably did not include sex.Footnote 59 In contrast, the two regulations adopted in 2024 will make it obligatory not only to collect API data for every flight entering the EU as well as selected intra-EU flights but also to include sex in API data.Footnote 60 However, the precise date when states must fully comply with the two new regulations is not yet set and will depend on when the technical router for the data transfer, which is established by eu-LISA, is completed.Footnote 61
Sex was likely included in the new API Regulations because this would align with existing practices of EU member states and international standards.Footnote 62 While sex is not among the data points listed in the 2004 API Directive, several EU member states already include genderFootnote 63 in their API data collection.Footnote 64 Moreover, guidance issued by the ICAO, the World Customs Organization and the International Air Transport Association recommend states request sex information as part of API.Footnote 65 Additionally, Annex 9 to the Convention on International Civil Aviation (Chicago Convention), which has 193 contracting states, notes that states may demand that airline companies share the machine-readable information displayed on passports, which, notably, includes sex markers.Footnote 66 Thus, the inclusion of sex as a mandatory data point in the new EU API framework reflects the rather established practice of requesting sex information in API data collection.
During the public evaluation of the 2004 API Directive, at least two anonymous responses suggested the treatment of sex as a mandatory API data point.Footnote 67 However, once the new API Regulations were drafted, doubt was expressed about including sex in API data. The German delegation raised the following questions: “[r]egarding the data category ‘sex’ mentioned under (b), the question arises whether the required travel documents always contain information on sex? If not, what is the basis for this information? Why is it required?”Footnote 68
While these questions were left unanswered and sex remained a data point in the new API Regulations, they highlight the role of travel documents, such as passports, in collecting sex information as part of API data. Indeed, the new API Regulations explicitly hold that specific information, such as a passenger’s sex and nationality, shall normally be collected through automated means that capture the machine-readable data on the passenger’s travel document.Footnote 69 Most API data is thus obtained by scanning MRPs during check-in, which makes sex markers on passports instruments of border security and migration control.
One more feature of the new API Regulations is worth mentioning. API data, including sex information, will feed into the EU’s newly established central repository for reporting and statistics (CRRS) to create anonymized travel statistics.Footnote 70 These statistics are used for understanding migratory patterns and determining priorities for future border management. Interestingly, at least two amendments proposed in the European Parliament suggested the removal of sex as a statistical element in API data.Footnote 71 However, these amendments were rejected, leaving sex as part of the statistics compiled through API data collection. Thus, sex markers on passports also serve as a biopolitical tool of big data collection, which assists governments in monitoring, determining, and regulating future risky migrants.
II. Classifying Risky Migrants Through Passenger Name Records
Another border security measure in the EU that relies on sex markers on passports is the collection of PNR. Similar to API, PNR is a set of information about passengers that airline companies share with border authorities prior to the arrival of an aircraft. While API includes data displayed on travel documents and limited flight details, PNR consists usually of any information about travelers that airline companies collect for their own purposes, such as advertisement strategies. What is particular about PNR in the EU is that it integrates API data, which means that if sex information is collected for API data purposes, it becomes also part of PNR data.Footnote 72 Once the two new API Regulations start to apply to relevant EU member states, sex information will always be part of their PNR data collection.
The history of PNR data resembles that of API data. The collection of PNR data was driven by the US in the direct aftermath of 9/11 and is by now fully endorsed as a security strategy by the UN Security Council.Footnote 73 In the EU, PNR data collection became mandatory for SchengenFootnote 74 states with the adoption of the PNR Directive in 2016.Footnote 75 According to Didier Bigo and Stefan Salomon, the PNR Directive followed not only the security concerns of the US but was “born also from the EU’s preoccupations with irregular migration,”Footnote 76 highlighting the use of PNR data for regulating migration in addition to tackling security concerns. The PNR Directive defines the PNR data that EU member states may collect, which includes personal data, such as names, addresses, contact information, and payment details, and flight details such as itinerary, frequent flyer status, and baggage. It notably includes API data and specifically lists “gender” and “nationality” as part of the API data that feeds into PNR data.Footnote 77
Once obtained, PNR data can be assessed to tackle terrorism and serious crime through two means. First, it may be cross-checked with existing databases.Footnote 78 Here, sex and nationality, both API data points obtained through the scanning of passports, can serve as filters for searches in these databases. Second, PNR data may be processed against predetermined criteria to identify “persons who were unsuspected of involvement in terrorist offences or serious crime prior to such an assessment and who should be subject to further examination by the competent authorities.”Footnote 79 This constitutes a predictive risk assessment where pre-determined criteria are used to calculate the riskiness of individuals. No provision in the PNR Directive explicitly stops states from using sex or nationality in PNR-related risk assessments, making it likely that states rely on these two data points in their assessments.Footnote 80
The legality of the PNR Directive has been evaluated by the Court of Justice of the European Union (CJEU) in Ligue des droits humains in 2022. A crucial outcome of the CJEU’s judgment in this case was that algorithms used for assessing PNR data against predetermined criteria cannot rely on “artificial intelligence technology in self-learning systems (‘machine learning’).”Footnote 81 The opacity of such systems could make it impossible for a human to review the algorithmic assessment,Footnote 82 which is required by the PNR Directive.Footnote 83 This means that algorithms used in PNR-related assessments cannot be based on a “black box.”
The CJEU further emphasized in Ligue des droits humains that the PNR Directive contains several non-discrimination provisions. It, for instance, specifies in Article 6 that assessments based on predetermined criteria must not be carried out in a discriminatory manner and that the predetermined criteria must be proportionate, specific, and regularly reviewed.Footnote 84 Furthermore, predetermined criteria cannot “be based on a person’s race or ethnic origin, political opinions, religion or philosophical beliefs, trade union membership, health, sexual life or sexual orientation.”Footnote 85
Sex and nationality are notably not listed in Article 6 of the PNR Directive. The exclusion of nationality from the list makes evident that race discrimination has been conceptualized under EU law in a way that “affords EU Member States significant leeway in discriminating against non-nationals simply because of their immigration status.”Footnote 86 In addition, some commentators, including the Fundamental Rights Agency, are skeptical that the PNR Directive effectively protects against discrimination on the grounds it does enumerate. The reason for their skepticism is that even without relying on the sensitive data listed in Article 6, the outcome of the risk assessments can in effect discriminate against groups on the basis of, for example, race, religion, or sexual orientation.Footnote 87 Indeed, while the risk scores assigned to different nationalities in PNR assessments remain confidential, it is likely that dynamics similar to those underlying the negative and positive visa lists are at play. Thus, using nationality as a risk input factor is likely to create indirect race discrimination by de facto deeming more Black and Brown people as “risky” than White persons. This would precisely demonstrate Achiume’s claim that nationality works as a proxy for race.Footnote 88
In fact, the CJEU recognized in Ligue des droits humains that assessments of PNR data against predetermined criteria must not result in indirect discrimination on the basis of, for example, race or ethnicity.Footnote 89 This recognition opens the door to challenge the use of nationality to calculate the riskiness of passengers through PNR data in the EU. However, as Henrik Palmer Olsen and Cornelius Wiesener point out, for indirect race discrimination to be revealed, race-disaggregated statistics on the “hits” in assessments enabled through PNR data must be made public,Footnote 90 which authorities seem unwilling to do.
The possibility to challenge the use of sex as an input factor in PNR-enabled risk assessments seems stronger because Article 21(1) of the Charter of Fundamental Rights (CFR), which the PNR Directive refers to,Footnote 91 lists sex as a non-discrimination ground.Footnote 92 Relying on sex in PNR-related assessments would result in differential treatment on the basis of sex, which seems contrary to Article 21(1) of the CFR. However, my analysis of the VIS below shows that, despite the prohibition of direct sex discrimination under EU law, risk assessments in the EU regularly rely on sex as an input factor. This is also suggested by research on an EU pilot border management project, iBorderCtrl, which takes gender into account in algorithmic risk assessments.Footnote 93 It therefore appears that Matthew Longo was correct in arguing that, in risk assessments at borders, “discrimination is a feature, not a bug—it is the essential way by which the good traffic is distinguished from the bad.”Footnote 94
III. Governing Mobility by Relying on Sex Markers in the Visa Information System
After having discussed two security measures used in EU border controls, my attention turns to the role of gender markers on passports in Schengen visa applications. While focusing on the functioning of the VIS, part of my analysis is also relevant for ETIAS, whose risk assessment rules are similar to the ones for the VIS. Once it becomes operational in approximately late 2026, ETIAS will record travel authorizations of third-country nationals who are exempted from Schengen visa. As explained below, both the VIS and ETIAS record the sex of travelers, which they use to calculate their riskiness. I will first analyze how the reliance on sex for risk assessments under the VIS and ETIAS helps to detect “unwanted” racialized migrants, before turning to the specific implications for gender non-conforming individuals.
1. The Racialized Traveller Through Sex Information in the Visa Information System
The VIS was formally established as a system to exchange data on visas among EU member states in 2004.Footnote 95 Its purposes and scope were defined in 2008, and it became operational in 2011.Footnote 96 However, the VIS is currently being reformed as part of the process of fortifying the EU external borders through the EU interoperability framework. In the context of this reform, the VIS Recast Regulation introduced significant amendments to the VIS, but the launch of VIS operations under this Regulation has been postponed until 2027.Footnote 97
The VIS records the sex of individuals applying for a visa to a SchengenFootnote 98 state.Footnote 99 Once recorded in the VIS, sex information can be used as an element in background checks that are currently being expanded through the VIS’s reform. When the new VIS system becomes operational, sex will feature in the information used to cross-check the other five databases that are part of the interoperability framework as well as the European Union Agency for Law Enforcement Cooperation (Europol) and International Criminal Police Organization (Interpol) databases.Footnote 100 In this process, sex works as a filter to find matches, similar to the way API and PNR data is cross-checked against existing databases. In addition, applications for short-stay visas for less than three months will be screened against specific risk indicators, which “shall be applied as an algorithm enabling profiling.”Footnote 101 Here, sex is one of the data elements that can be compared against risk indicators, alongside, for example, a person’s nationality, name, date of birth, residence, occupation, member state of destination and first entry, and purpose of travel.Footnote 102 As was made explicit in the VIS Recast Regulation, this means that the relevant risk indicators can be based on information on sex in combination with other data.Footnote 103
It is thus clear that Schengen states can take sex into account when conducting risk assessments as part of the decision to issue or refuse a visa. It is interesting though that the VIS Recast Regulation specifically holds that risk indicators cannot be based solely on the sex or age of a person.Footnote 104 This means that sex or age must always be combined with other factors when assessing people’s riskiness. It seems that the main reason behind the existence of this provision was the fear that using sex or age as the sole determinant to decide on a person’s worthiness to receive a Schengen visa could constitute direct sex or age discrimination. This becomes evident when analyzing the procedural history of the regulation establishing ETIAS (ETIAS Regulation), which contains the same list of data that can feed into risk indicators as the VIS Recast Regulation.Footnote 105
The screening of ETIAS applications against risk indicators has been widely criticized by scholars and human rights institutions based on concerns over data protection and discrimination.Footnote 106 Albeit not the core concern of skeptics of ETIAS, the use of sex and age in risk indicators also raised debate during negotiations on the ETIAS Regulation in 2017. Some members of the European Parliament proposed including sex and age in the list of characteristics that can never be revealed by risk indicators, such as a person’s color, race, religion, disability, or sexual orientation. This suggestion was justified by reference to Article 21 of the CFR, which, as discussed above, lists sex and age as non-discrimination grounds.Footnote 107 The Council took this suggestion on board and also proposed including sex and age in the list of prohibited input factors in a trilogue with the Commission and Parliament. The compromise of permitting the use of sex and age as risk indicators under the ETIAS system, while clarifying that they cannot serve as the sole factors, appears to have been introduced during this trilogue.Footnote 108
It therefore seems that concerns about potential direct discrimination, contrary to primary EU law, prompted the specification that sex and age cannot serve as the sole determinants in risk assessments under ETIAS and the VIS. The assumption was that, if sex and age are combined with other factors, such as a person’s occupation, relying on them is no longer discriminatory. However, as noted by Charly Derave, Nathan Genicot, and Nina Hetmanska, this ignores the concept of intersectional forms of discrimination, as it does not prevent sex or age from creating a disadvantage in intersection with other statuses.Footnote 109 For instance, while assigning all men a higher risk score than all women is prohibited under rules governing ETIAS and the VIS, men of certain nationalities may still be classified in a higher risk category than men of other nationalities or than women of the same nationality. In fact, this was precisely the outcome of an algorithmic risk assessment used by the Netherlands for issuing short-stay Schengen visas, which automatically flagged all men from Suriname for intensive scrutiny.Footnote 110 Thus, the risk assessments undertaken under the regulations governing the VIS and ETIAS show that, despite sometimes being described as a “buzzword,”Footnote 111 intersectionality is far from being effectively enshrined in EU law.Footnote 112
In addition to not considering intersectional forms of discrimination, the ETIAS Regulation and VIS Recast Regulation have been criticized for not effectively protecting against indirect discrimination. Scholars have noted concern that the screening of ETIAS applications against risk indicators results in indirect discrimination, especially indirect race discrimination through the reliance on nationality as an input factor.Footnote 113 This concern has led the Fundamental Rights Agency to recommend postponing the use of risk indicators until a test phase proved the absence of discriminatory profiling, while the European Data Protection Supervisor urged subjecting the ETIAS screening rules to a comprehensive fundamental rights impact assessment.Footnote 114 Likewise, several parliamentarians proposed removing the possibility of assessing ETIAS applications against risk indicators altogether,Footnote 115 while others proposed specifying that the ETIAS system shall not result in direct and indirect discrimination.Footnote 116 However, the final versions of the ETIAS Regulation and the VIS Recast Regulation do not reflect these concerns.
The possibility of engendering race discrimination by relying on nationality in the risk indicators used for assessing applications under the VIS highlights that the racialized nature of the Schengen visa system is established at two levels. At the first level, individuals are separated into those that need visas and those exempted from the visa requirement based on racial lines by the above-mentioned positive and negative lists. At the second level, the worthiness of those who are covered by the negative list and therefore need a Schengen visa is calculated by relying on nationality, which works as a proxy for race. At this level, sex functions in intersection with nationality as an element to establish who is risky and therefore racialized through the risk assessments under the VIS. The filtering of visa applications based on sex in the VIS has therefore a racializing function.
2. Gender Non-Conformity as an Indicator for Suspicion
The reliance on sex in VIS-related risk assessments also creates specific repercussions for trans and non-binary persons. Specifically, the use of sex as a risk indicator under the VIS assumes that sex is a stable and binary characteristic. This assumption creates situations in which gender–nonconforming individuals are flagged as suspicious due to discrepancies in gender markers across different documents.
For instance, a mismatch can occur when the sex recorded in the VIS does not correspond to the sex marker on a person’s passports. Rather than relying on the passport, the VIS extracts sex information from the Schengen visa application form, which lists only the binary categories male and female.Footnote 117 As a result, persons with a gender marker other than F or M on their passport might tick either the F or the M box on the application form, despite this marker not corresponding to their passport. However, passports or other travel documents must be presented together with the application form.Footnote 118 As a result, the Schengen visa system treats non-binary persons as abnormalities and creates situations where a person’s sex marker on their visa application form might differ from their passport, which can likely lead to additional verification or delays in processing the application.
Although no relevant cases involving the Schengen visa application system are yet public, experiences from other jurisdictions indicate that treating individuals whose gender is other than female or male as an abnormality within visa systems can lead to adverse consequences for them. For instance, an Indian trans person reported in 2015 that her US visa application was “put on hold” due to a discrepancy in gender markers between her passport and visa application. On her passport, she was recognized with a T (standing for “transgender”), while she ticked F on the visa application form as the form only listed binary categories. This mismatch made her illegible to the binary gender model enshrined in the US visa application system, which resulted in the failure to process her application.Footnote 119 Given that the Schengen visa system is similarly based on the gender binary, gender non-conforming persons could easily have comparable experiences when applying for visas to Schengen states.
Situations like the one above highlights the broader problem that gender non-conformity is frequently treated as a ground for suspicion. In the aftermath of 9/11, certain institutions, such as the U.S. Department of Homeland Security, suggested that terrorists—who were assumed to always be men—could disguise themselves as women.Footnote 120 This turned gender non-conformity into an indicator of identity fraud, especially if it was expressed by non-White individuals. As Toby Beauchamp argues, trans persons who conform to normative presentations of race pass more easily as “properly gendered and safe,” while bodies transgressing those presentations are more commonly read as suspicious. After all, “[d]ominant notions of what constitutes proper feminine or masculine behavior are grounded in ideals of whiteness, class privilege, and compulsory heterosexuality.”Footnote 121
The intersection of race and gender in the construction of transgressive bodies is also made visible through data on the effects of gender markers on IDs. For instance, the LGBTI III Survey conducted by the EU Fundamental Rights Agency found that among the 31,232 trans, non-binary, and gender–diverse respondents, 17% reported having experienced discrimination when presenting an official document indicating their legal sex in the past year. Yet, rates were higher among those belonging to religious, racial, ethnic, or migrant minorities (20–22%) and those who were asylum seekers and refugees (31%).Footnote 122 This indicates that gender non-conformity is especially policed in non-White individuals and migrants.
The scrutiny experienced by gender non-conforming persons in security checks has generally heightened with the expansion of border control measures after 9/11.Footnote 123 Trans persons are especially affected by measures taken against identity fraud because their identity documents often reflect different gender markers. For instance, the US Real ID Act was passed in 2005 with the aim to clamp down on undocumented migration by securitizing and standardizing identity papers. Unsurprisingly, this has hindered access to official documents and state services for undocumented migrants, racialized communities, and poor people. In addition, it has increased the risk of being flagged for trans individuals because state registries often record their gender inconsistently, as policies for changing gender markers differ across US state bodies.Footnote 124
The new interoperability framework in the EU risks enshrining a similar model in which gender non-conformity becomes an indicator for fraud. Once the framework starts operating, a file created in the VIS is cross-checked against other databases connected through the interoperability framework using Multiple-Identity Detection. If the file appears to match an existing profile in another database, but with some differences in personal data, it is flagged as a “yellow link” and referred to national authorities for manual verification.Footnote 125 This could affect trans persons who have changed their legal gender and/or name, as their profiles across different databases match except for the gender and/or name entry.
For instance, if a trans man who was previously recorded as female under his deadname on his long-term visa in the VIS applies as a man with his current name for a travel authorization under ETIAS, he might be flagged because his biometric data in the VIS and ETIAS is the same, while his identity data differs. His file would consequently be sent to national authorities for further verification. Even though national authorities can rectify any data discrepancies, trans persons might nevertheless initially be flagged as in need of extra verification. This treats trans persons as potential fraudsters, which is something that scholars and activists have documented in various other contexts.Footnote 126
Thus, recording sex in the VIS, and ETIAS, is one tool among many to identify migrants and those who are subjected to extra scrutiny. The extent to which the assessments under the revised VIS and ETIAS will be steered by artificial intelligence and self-learning systems remains to be seen.Footnote 127 Based on reports by eu-LISA and the EU Commission, it seems likely that they will, eventually, rely on machine-learning techniques for calculating the riskiness of travelers.Footnote 128 If this occurs and the CJEU is called upon to address the issue, it will be interesting to see whether the Court reaches a decision similar to Ligue des droits humains, where it held that PNR data may not be assessed against predetermined criteria using machine-learning algorithms. In addition to the Court, the EU Artificial Intelligence Act, adopted in 2024—and whose application to migration control deserves further discussion elsewhere—may also place limits on risk assessments carried out under the VIS and ETIAS. However, it will not apply to the large–scale IT systems for borders and security until December 31, 2030.Footnote 129
E. Beyond the Gender Binary or a New Form of Biometric Control?
Given the various ways gender markers on passports are increasingly used in risk assessments in the EU, it is rather unlikely that states will soon move toward abolishing them. Nevertheless, it is worth asking what the future of these markers should be. Would abolishing gender markers on passports make borders less gendered and racialized?
As my analysis shows, passports and information on passports, such as gender markers, are no longer the ultimate tools on which decisions on a person’s right to cross a border are taken. Instead, states increasingly draw on an assemblage of various types of information, including travel history, information on occupation, and banking details. According to Dimitri van den Meerssche, legal identity categories, such as citizenship, are increasingly deprioritized in algorithmic risk assessments and only function “as one of many in the ‘bundle of vectors’ assembled for decision-making purposes.”Footnote 130 Similarly, Dennis Broeders and Huub Dijstelbloem argue that, with the rise of big data that informs visa decisions, “some categorisations now move away to some extent from the rather blunt categories of nationality towards being based on characteristics of (aggregated) individuals.”Footnote 131 While it is only partly true that categories like nationality lose relevance in determining the “(un)wanted” migrant, as demonstrated by the positive and negative EU visa lists, people’s legal identity is indeed no longer only made up of state-governed characteristics, such as gender and nationality. Rather, legal rights, such as the right to enter a state, are increasingly also shaped by digital traces, often described as a “digital footprint,” which one leaves on the web, such as through social media use, booking flights, and financial transactions.Footnote 132
Yet, the reliance on data, such as digital footprints, for determining a person’s rights is not gender- or race-neutral. Rather, gender and race and many other inequalities, such as class, can shape people’s digital footprints by affecting their behavior online, such as which content they consume, and material access to digital services like e-banking and the purchase of flights. Indeed, companies sometimes infer a (binary) gender and race from a person’s digital footprint for targeted advertisement.Footnote 133 Data can therefore be gendered and racialized, even without relying on formally attributed gender and race labels, which makes decisions taken based on this data, such as whether to subject somebody to further scrutiny at borders, far from gender– or race-neutral.
In addition to big data, states increasingly rely on biometric technology, such as fingerprint, facial, and iris recognition, for governing people’s movement. This becomes evident in ICAO’s recent announcement to launch a pilot of a Digital Travel Credential (DTC) system, which aims to replace traditional boarding passes and check-in processes with biometric and digital identity technology. While DTC systems will store passports digitally, the ICAO plans to use facial recognition to identify people at airports and borders.Footnote 134
In fact, the rise of biometric technologies such as facial recognition has led to questions about the continued relevance of gender markers for identification. For instance, the above-mentioned review of gender markers on passports presented by New Zealand to the ICAO in 2012 notes that “biodata,” which includes gender information, may become less relevant for confirming identity with the advancement in biometric identification methods.Footnote 135 Similarly, the UN Statistics Division held in 2024 that “[w]ith improvements in biometrics, it may not be necessary to include sex or gender information on identification documents at all.”Footnote 136 Biometrics is thus sometimes described as an alternative to using gender markers on passports as tools of identification.
Yet, once again, border security does not necessarily become less gendered or racialized if certain data on passports, such as gender markers, is increasingly replaced or, at least, complemented by biometric data. As Achiume points out, border technology is far from race-neutral, despite often carrying the guise of objectivity and neutrality.Footnote 137 Rather, it is racialized not only because it is often trained on racially non-representative data, such as in facial recognition software, but also because its very purpose is to protect the border from “unwanted” migrants—a concept that has always been shaped by racial and gender norms.
For instance, facial recognition software has been criticized for reproducing gender and race biases. It is widely known that facial recognition software often leads to a higher error rate for people of color, and especially women of color, as it is mostly trained on the data of White people and men.Footnote 138 In addition, facial recognition systems often assign trans persons and women of color incorrect gender classifications because they rely on cisnormative and White facial norms, particularly in relation to features such as lips, eyes (including make-up) and cheeks.Footnote 139 Moreover, persons with a gender identity other than female or male are almost always misclassified by facial recognition systems because the relevant software codes gender as a binary—female or male.Footnote 140 A racialized form of the gender binary is thus encoded into facial recognition software.Footnote 141
Given the higher numbers of gender misclassification of trans persons, especially non-binary persons, and persons of color by facial recognition technology, they are flagged more often as suspicious by these technologies. This treats trans persons and persons of color as deceptive or potential terrorists, similar to discrepancies in gender data across different databases discussed previously. In discussions about surveillance through biometric technologies, such as facial recognition, trans people have argued that “[j]ust by virtue of being trans, we are perceived as people who can only be terrorists and thieves.”Footnote 142 Similarly, Nicholas Clarkson holds that “[b]iometrics are thoroughly embedded in projects of racializing surveillance and other forms of bodily normativization.”Footnote 143 Biometric technology is therefore not less cisnormative or racialized than official documents.
In fact, fixing the gender and race biases in data-driven border technologies, such as facial recognition software, does not mitigate the fact that the very purpose of these technologies is to tighten racial borders. As Jens Theilen and others note, efforts to diversify data sets “can also distract from the fact that certain technologies should, from a feminist standpoint, be opposed entirely rather than merely being made more inclusive and accurate.”Footnote 144 Even if it was possible to ensure biometric border technologies rely on non–biased data, they would nevertheless intensify what Simone Browne calls “racializing surveillance” precisely because they have often developed as a means to reify racial boundaries.Footnote 145 Similarly, as Os Keyes argues, gender recognition technology that avoids the misclassification of trans and non–binary individuals would still reinforce the idea that gender is linked to physiological characteristics and can therefore serve as an element of identification.Footnote 146 Thus, well-intended efforts to fix racial and cisnormative biases in datasets underpinning biometric technology, such as facial recognition, could paradoxically bolster the efficiency of these technologies for the policing of racialized and gender non-conforming bodies.
In addition, it could be even harder to decode cisnormativity in biometric technology than state registries and official documents. Once facial recognition software has registered a facial image, this snapshot often becomes set in stone and difficult to replace with other images. This poses significant obstacles to trans persons whose facial features often change when transitioning. Instead, in jurisdictions where legal gender recognition is easily accessible—which, admittedly, is the vast minority of states—the gender marker on passports can be altered through a more or less self–determined process. Holding passports from such jurisdictions could, in fact, even serve as a tool to fight biological determinism enshrined in biometric technologies.
For instance, a biometric tool widely used at airports nowadays is body scanners. When individuals enter these scanners, security guards usually classify the passenger into a binary gender category based on their own perception by pressing a blue button (male) or a pink button (female). On the basis of this gender classification, trans persons are often flagged for displaying some sort of “abnormality” when certain body parts or prosthetics do not align with the gender norms associated with the pressed button.Footnote 147 When this happens, the passengers concerned are often subjected to additional security checks, including body searches. In such moments, holding a passport with a gender marker that corresponds to one’s gender identity can help to challenge the biometrically assigned gender classification, such as by requesting to be pat down by an officer with a corresponding gender identity. Indeed, the British Passport Office recognized in a review of gender markers on passports in 2014 that these markers can enable “transgendered persons to use their passport as evidence of their acquired gender.”Footnote 148
The example of body scanners at airports shows that states’ gender classification marking on passports is no longer the only gender classification system mobilized in border crossings. Instead, trans persons’ identities are governed by multiple, often competing gender classification systems, including biometric technology, the security officers’ perception, their state’s passports, and their own self-identification. Abolishing one of these competing classifications, such as gender marker on passports, would transfer the regulatory power over gender assignments to the other systems. While this could reduce misclassification by supporting gender self–identification, it could possibly also enhance the power of biometric technology or security officers’ perception to box people into gender categories. The gender binary would not lose its relevance if it was no longer certified through passports; it would continue to work through other classificatory systems until we live in a hierarchically gendered world.
However, this does not mean that abolishing gender markers on passports is not a valuable pathway. In fact, it could eliminate one data point among the many data points which states rely on to protect their racialized borders and police bodies based on cisnormativity. In the past, certain passports, like the US passport used in 1818 or the British passport from 1915, noted the category of “complexion.”Footnote 149 The elimination of race categories like “complexion” from passports was certainly a positive step to reduce the overt racism enabled through passports, but scholars like Achiume have effectively shown that exclusions based on passports have not become race–neutral. Instead, race continues to work through direct and indirect means, including through exclusions justified on the basis of nationality and various “neutral” variables used in risk assessments leading to racialized outcomes. In line with what Critical Race Theory has taught us about race-neutrality,Footnote 150 the racial border has not disappeared by removing the category “race” from passports.
Similarly, given the fact that the gender binary has pervaded social structures and institutions, it would not fade with the elimination of binary gender categories in the law. Certainly, states have often oppressively imposed the gender binary, like racism, through various laws, policies and institutions, including those that oblige the notification of gender on passports. Given their authoritative force, states have normalized cisnormativity and the policing of gender non-conformity. In turn, removing gender markers from identity documents, such as passports, could also carry symbolic force that initiates a broader reconsideration of gender norms.
Nevertheless, state law is not the only normative system in which humans are embedded and where the gender binary continues to hold its grip. Rather, data, algorithms, and biometrics, which increasingly determine our rights and duties, such as the right to enter a state, are equally gendered—and racialized. Data, algorithms, and biometric systems are certainly not fully autonomous agents; behind them are persons, such as those verifying the results of risk assessments carried out under the VIS or with PNR data, those mining data from social media platforms, and those programming the algorithms, even when they evolve into self-learning creatures. Yet, the persons behind the technological tools that determine our rights are not always state agents but increasingly private companies. This is demonstrated by the fact that several elements of the EU’s interoperability framework are developed by the private sector, such as the Biometric Matching System being created by the company Idemia.Footnote 151 Thus, the oppressive force of the gender binary—and of racism—stems not only from the state but also from non-state structures and actors.
F. Conclusions
Analyzing the origins and current functions of gender markers on passports highlights that these markers are not an accident of cisnormative history. Legal categories have always existed for specific purposes. Marking gender in the form of an F, M, or X on passports has been, inter alia, a tool of border governance which helps identify people and trace bodies that are deemed risky. As the notion of “the unwanted” or “the risky” has always been informed by race and gender norms, gender markers become indirectly a tool of securing the gendered and racialized border.
With the expansion of risk assessments undertaken under EU law regulating API, PNR, and VIS, which use gender as an input factor, it becomes less and less probable that gender markers on passports will be abolished. As my analysis shows, gender markers have not always existed on passports, but systematically emerged since reliance on machines to read travel documents increased. The continued digitalization of border controls further intensifies the reliance on narrowly boxed categories that can be encoded and analyzed by machines and algorithms to detect travelers of interest. Travelers are dissected into various sub-parts of their ascribed identity to analyze their riskiness. Gender is one of these markers that, in combination with other markers, such as nationality, determine who will be placed under surveillance or denied entry to a state.
The precise role artificial intelligence will play in constructing travelers’ risk profiles in the EU is not yet clear, but it seems likely that algorithms will, at some point, further dissect the input factors, combine them with other data, and reassemble them into profiles whose components are hardly graspable by humans. In this process, gender might no longer have the shape and form of clearly demarcated boxes on passports, but gender would still inform the algorithms and the definition of risky migrants through reliance on gendered data about a gendered world. The gender binary has become so deeply ingrained into social structures, technology, and data that reducing states’ gendering power would not result in a gender-free world. While eliminating gender markers from passports could indeed abolish one data point on the basis of which gender non-conforming individuals are policed and the risky traveler defined, the gendered and racial border would not disappear. For this, we need far wider systematic dismantling of global racism and transphobia.
Acknowledgments
This Article has benefited greatly from the feedback and support of many individuals. I am grateful to Joy Magomba for excellent research assistance; to Miriam Fahimi and Alpa Parmar for their invaluable comments; to the two anonymous reviewers for their deep engagement and thorough feedback; to the other participants in the Sofia workshop for their comments; and to the editors for their patience and commitment to this issue.
Competing Interests
The author declares no competing interests.
Funding Statement
The author received no specific funding for this Article.
Open access
