A study on the use of new technologies in the identification and search for missing or deceased persons in migration

DECEMBER 2025 © EuroMed Rights 2025 A STUDY ON THE USE OF NEW TECHNOLOGIES IN THE IDENTIFICATION AND SEARCH FOR MISSING OR DECEASED PERSONS IN MIGRATION. Written by Marie Godin, Emma Johnston and Giorgia Mirto December 2025 EuroMed Rights Vestergade 16, 2nd floor DK-1456 Copenhagen K Denmark Tel: +45 32 64 17 00 [email protected] www.euromedrights.org Bibliographic Information Title: A study on the use of new technologies in the identification and search for missing or deceased persons in migration Authors: Marie Godin, Emma Johnston and Giorgia Mirto Date of Publication: December 2025 Pages: 58 Original Language: English Marie Godin, is a Lecturer in Human Geography at the University of Leicester in the School of Geography, Geology, and the Environment. Additionally, she is a research affiliate at the Centre on Migration, Policy, and Society (COMPAS) at the University of Oxford. She is also an associate editor for the Journal of Migration Studies, published by Oxford University Press (OUP). She has, with colleagues, just edited a Special Issue in the Journal of Ethnic and Migration Studies entitled “Digital technologies and migration: behind, beyond and around the black box” (2025). The Gerda Henkel Foundation supported her time during this research. E: Marie[email protected]x.ac.uk Emma Johnston is a Senior Lecturer in Forensic Biology at De Montfort University. As a forensic scientist in the field of DNA profiling, she worked on cases in the UK criminal justice system as well as with the Forensic Anthropology Foundation of Guatemala. Her PhD focussed on “The Role of Forensic Science in Identifying Missing Migrants in Europe”. She is a member of the COST Action CA22106 Migrant Disaster Victim Identification (MDVI). She is currently working on the ESRCfunded project “Trust in forensic science evidence in the criminal justice system: The experience of marginalised groups”. E: [email protected] Giorgia Mirto is a PhD Candidate at the Department of Anthropology at Columbia University. In her PhD project, she examines the treatment of bodies, relatives, material remains, and reliquary traces in the aftermath of migrant maritime disasters, and how these shape the struggle over the incorporation of the body and person of the unknown migrant in Sicily. For over 15 years, Giorgia has been deeply involved in various research projects about Italian protocols for managing, burying, and identifying migrants' bodies recovered from the Mediterranean Sea. Beyond academia, Giorgia's activist engagement contributes to a wide network of actors aiming at supporting family members of the disappeared and their collectives who are seeking truth and justice on both sides of the Mediterranean. E: [email protected] AKNOWLEDGMENTS We extend our sincere thanks to the interviewees who generously shared their time and expertise with us. We are particularly grateful to Filippo Furi for his guidance throughout the process, for connecting us with key interlocutors, and for bringing the team together. We also want to thank Carolina Sanchez Boe (Adjunct Lecturer, Brown University, Paris) for her attentive reading. The authors are solely responsible for the data and the content of the paper. We also wish to acknowledge EuroMed Rights for their trust, patience, and the very insightful comments they provided on the draft report. Grammarly was used to ensure grammatical accuracy throughout this report. This research was possible thanks to the support of the European Artificial Intelligence & Society Fund. The sole responsibility for the content lies with the authors and the content may not necessarily reflect the positions of NEF, or the Partner Foundations. TABLE OF CONTENTS Executive summary General recommendations List of acronyms I. Introduction II. Forensic approaches to the search for missing migrants and the identification of deceased migrants 1. The case of missing migrants in the mediterranean represents a sui generis scenario III. Critically assessing new technologies in the search for missing and deceased migrants IV. Mapping the tools for search and identification 1. Complex social network analysis (csna) 2. Open source investigation (osint) and social media intelligence (socmint) 3. Dvi & mdvi 4. Dna 5. Rapid dna 6. i-Familia 7. Dna & ai 8. Fingerprints 9. Facial identification 10. Age estimation 11. Data & databases 12. Access now's digital secuity helpline 13. Social media platforms as spaces for memory, commemoration, resistance and mourning V. Ethical considerations 1. The lack of a regulatory framework, more than a technical discussion 2. Regulations and guiding principles around ai tools 3. The rights of families of the missing and the use of new technologies VI. Conclusion Bibliography Appendices Interview grid List of participants 5 6 10 11 15 17 19 21 21 24 28 30 30 31 32 33 33 36 36 41 42 44 44 45 46 48 49 55 55 56 EXECUTIVE SUMMARY In recent years, “new technologies” have become central in the development and evolution of security and migration control systems, within the European Union (EU) and globally. This concerns both the control and biometric tracing of bodies, as well as the development of technologies for profiling and tracking virtual identities, often linked to the use of artificial intelligence (AI). Against this backdrop, thousands of migrants have died on dangerous journeys to Europe. We do not know who the vast majority of these individuals are. This leaves friends and family in a state of limbo, not knowing the fate of their loved ones. In this report, we examine the potential implications of the application of new technologies in the specific context of border deaths: the search for missing migrants along migration routes and the forensic identification of bodies of border deaths. Here, the deployment of new technologies has important consequences for the families of missing persons seeking answers from European authorities regarding the fate of their relatives, and ultimately, their search for justice. Through desk-based research and 30 qualitative interviews conducted with forensic experts, institutional representatives, civil society organisations, legal specialists, and family or community members we were able to map and critically evaluate new technologies in the search for the missing and the identification of deceased migrants. To better explore the issue of use of new technologies in the identification and search for missing or deceased persons in migration, we begin by examining the role of forensic experts, who mediate between the demands of families and institutions, and the approaches they take. We explore different genealogies within the field of humanitarian forensic action, such as the “Latin American Model”, the work of the International Committee of the Red Cross (ICRC), counter-forensics and citizen-led forensics approach. We note that the emerging development of AI-based forensic technologies aimed at locating and identifying migrants who have died or gone missing in the Mediterranean is situated within a complex entanglement of the forensic, the humanitarian, and the political. We caution that centring material traces and forensic identification risks overlooking the broader context of missing migrants and the role and agency of families in the search. Furthermore, the case of missing migrants in the Mediterranean constitutes a unique setting for this work as the fundamental trace essential to forensic work is often absent: the body itself. We explore, in detail, methodologies involving new technologies and artificial intelligence related to the search for the missing such as Complex Social Network Analysis (CSNA) and Open Source Intelligence (OSINT). As well as those more closely linked to forensic identification of the dead such as Disaster Victim Identification (DVI), DNA, fingerprints, facial recognition and databases, while acknowledging that this distinction (between search and identification) is not always possible or helpful. Despite significant attention to the potential of AI, these technologies have not yet achieved widespread adoption. Their application remains fragmented and localised, primarily because experimentation is confined to a limited number of actors and cases, with only minimal collaboration across different stakeholders. We then outline the vitally important ethical considerations surrounding the use of new technologies in the context of missing migrants. Regulations and guiding principles are essential to mitigate potential harms from the use of new technologies, and the rights of families of the missing must be considered at all times. We conclude that new technologies are neither good nor bad, rather, it is a question of use, governance, and ethics and present our key findings and recommendations: 5 GENERAL RECOMMENDATIONS Innovative technologies and methodologies can play a vital role in strengthening humanitarian responses to missing persons and deceased migrants. They can improve the technical work of identification while also supporting the families who continue to search for answers. However, their use must remain rooted in truth, justice and humanitarian principles: focused on restoring dignity and rights, rather than reinforcing systems of border control and migration management. For this reason, any technological intervention must be supported by clear regulatory and operational frameworks that protect the dignity of the deceased and missing, and ensure the safety and rights of families, survivors, and all actors involved in the process. While new technologies can enhance the effectiveness of search and identification efforts, their value depends on a broader shift in regulatory, ethical, and political approaches: Tools must be guided not only by technical performance, but also by the purposes for which they are developed and the contexts in which they are used. Currently, many national and international migration policies continue to prioritise restrictive approaches. In this policy context, the development of identification technologies is likely to strengthen the EU’s digital border infrastructure, further entrenching security-driven frameworks. Before moving to technical recommendations, it is therefore necessary to call for a structural reorganisation that places humanitarian, political and social priorities at the centre of action. First and foremost, there are no binding national or international obligations requiring authorities to search for missing migrants, identify the deceased, or provide continuous support to families. Existing international frameworks – such as the Global Compact (Objective 8), the Rabat Process (National Focal Points Network), and the Council of Europe’s Resolution 2569 (2024) – remain non-binding, and few States have translated them into sustained political or operational commitments. Practice on the ground is fragmented and shaped mainly by investigative approaches focused on establishing legal responsibility – i.e., identifying smugglers rather than border victims – undermining efforts to restore truth and dignity to families. As a result, identification processes depend heavily on local authorities' initiative, while families often rely on NGOs, community networks, and international organisations. Major incidents, such as shipwrecks in Italy and Greece, have prompted national authorities to activate exceptional identification mechanisms. However, despite their relative successes, these efforts have had limited influence on broader institutional practices and have not led to sector-wide scaling. As a result, they remain isolated and are exceptional cases. Therefore, expanding action beyond judicial frameworks – while ensuring data protection, privacy, and personal security – would enable stronger engagement from families, NGOs, and community actors. Such changes would support better coordination among stakeholders. They would formally recognise the essential work already carried out by actors outside official forensic systems, who contribute daily to searches and early identification despite operating without institutional acknowledgement. The recommendations that follow are grounded in this structural perspective. They address different actors and areas of action – including search, identification, and information management – and consider both existing technologies and how future systemic reforms could enable more humane, effective, and inclusive use of these tools. EUROMED RIGHTS - A study on the use of new technologies in the identification and search for missing or deceased persons in migration 1. Recommendations for State Actors The primary recommendation for state authorities is both ethical and political. It calls for introducing a normative framework that compels the European authorities to provide answers, truth, and justice to families seeking information about their loved ones. It is central to an approach that reconfigures the investigative system to ensure more effective search and identification of migrant persons, and the proper use of available tools (both old and new ones), within a framework that is non-discriminatory, non-criminalising, and protective of families, survivors, witnesses, and support actors. All technical and methodological recommendations should be understood within this ethical and political context. Specifically, identification procedures concern primarily the countries where bodies are recovered and received. This includes: The application of identification techniques. • The organisation, standardisation, archiving, and tracking of information (biographical data, material evidence, testimonies) to enable future search and identification. • Coordination among all actors involved, including investigators, law enforcement, medical examiners, municipal administrators, but also family members, survivors, migrant communities and civil society actors. It is crucial to implement and reorganise identification systems, using new technologies or refining existing tools, while always ensuring the protection of families and witnesses' rights. Standardising and systematising the collected information facilitates formal identification and the issuance of official death certificates. To strengthen the link between searches for missing persons and the identification of bodies, cooperation with the countries of origin of missing persons — or the countries where their families reside — is essential. This ensures adequate technical support, including the collection and transmission of DNA samples, and enables the safe, secure exchange of sensitive information. 2. Recommendations for Transnational Institutions Although the technical management of identification processes is primarily the responsibility of national authorities, transnational institutions (such as the African Union and the EU) should support practical collaborations to enhance family tracing and identification efforts. This support should be both structural – by helping to establish cooperation frameworks specifically designed to address the issue of missing migrants – and operational – by facilitating collaboration between states and engaging with humanitarian organisations and civil society actors, whose roles in this process should be further legitimised. 3. Recommendations for International Organisations International organisations can play a fundamental role in this landscape, given their technical expertise and comprehensive vision. They must maintain a central role in interstate diplomacy, supporting national authorities with methodologies and practices that protect families, and helping legitimise and protect the role of civil society organisations in these processes. 7 EUROMED RIGHTS - A study on the use of new technologies in the identification and search for missing or deceased persons in migration 4. Recommendations for AI Developers, Academic Institutions, Forensic Experts, CSOs, and funders The primary goal of identification procedures is to provide families with clear, accurate, and timely answers. While the development of new technologies and scientific advancements is valuable, these efforts must never take precedence over the needs of families, who rely on prompt and reliable information about their missing loved ones. Active participation of families, migrant communities, stakeholders, and civil society in the design, development, and use of new technologies is therefore strongly recommended. The implementation of research and identification technologies and methodologies is often tied to the work of research laboratories (private or public), funded by various international or national institutions. In that regard, we should emphasise the humanitarian and social dimension of research and identification activities, and separate research aimed at developing innovative tools for other purposes, particularly when associated with the control of human mobility and activities that restrict the rights of people on the move. Funding should be provided for the development of tools specific to the context of the search for and identification of missing migrants, rather than borrowing or adapting tools from other fields. Nowadays, the development of new identification technologies is primarily directed toward perfecting identification tools and currently prioritises the biological/biometric dimension (facial recognition, DNA comparisons, etc.) and the identification of bodies. This approach tends to increase the technical capacity of state authorities responsible for identification, while reducing the scope for intervention and participation of complementary (but essential to identification activities) actors and families. A family-centred approach should guide the development of any new technologydriven tool designed to improve the identification and search for missing migrants. Placing families at the heart of these processes ensures that their knowledge, needs, and lived experiences shape both the design and implementation of such tools, ultimately making them more effective, humane, and responsive to those most directly affected. Search and identification procedures for missing migrants – including those using new technologies – must be understandable and accessible to families and their communities. Translation, for instance, should not be limited to linguistic interpretation; although essential, it is not sufficient. As demonstrated in the evolution of established identification techniques such as DNA analysis, families cannot be passive recipients of results. They must understand and engage with the process to ensure meaningful outcomes. A scientifically valid identification that families do not accept is, ultimately, a failed process. Families are not only a means to achieve identification – they are also its purpose. The methods used must be transparent and must aim for the highest possible accuracy. Evidence shows that combining multiple identification methods often yields the most reliable results. Moreover, an effective system must work not only in exceptional or high-profile cases but also in the many everyday tragedies that characterise migration journeys. This requires flexible, context-sensitive approaches that adapt tools and methodologies to each situation. It is both ethically imperative and procedurally appropriate to prioritize the fastest and most effective techniques available, including AI technologies, in order to provide answers as soon as possible. Any tool used in the search and identification of migrants must take into account the profound uncertainty and emotional distress experienced by families, friends, and communities of the missing. The absence of certainty – or of a body to mourn – creates a “frozen grief” that can lead to prolonged anxiety and confusion. The humanitarian necessity to provide timely answers can come into tension with the development of cutting-edge forensic methods, which —also due to the limited pressure from families and the public to identify deceased migrants—may involve 9 long experimental phases. While scientific research is essential, it must not overshadow the moral and political responsibility to place the well-being and rights of migrants and their families at the centre of all efforts. It is therefore vital to implement the search and rapid response capabilities of a collaborative system based on a holistic forensic approach; at the same time, it is essential to develop mediumand long-term data archiving and management practices and techniques that allow families to continue their research over time. A responsible and ethical use of technology requires resisting the temptation to treat AI as an automatic or superior solution and instead prioritising holistic, context-sensitive approaches that integrate multiple methodologies, local knowledge, and the perspectives of families and communities. There is growing pressure within research and pilot projects to reference AI – even when it is not substantively necessary – because doing so is often perceived as increasing the likelihood of securing funding. This funding-driven incentive risks distorting priorities in migrant identification, leading to the development of technologically sophisticated solutions that may not align with the real needs of families, practitioners, or affected communities. It also creates an environment in which emerging tools are promoted and tested prematurely, despite limited accuracy or insufficient validation in highly sensitive contexts such as the identification of people who die at sea. Humanitarian technophilia, an enduring belief that technological innovation can inherently improve humanitarian action, is not a new phenomenon. However, the deployment of new technologies, including AI, in humanitarian contexts carries significant, well-documented risks. Such technologies can, in practice, generate new forms of vulnerability, exacerbate existing power asymmetries, and ultimately cause more harm than good. Concerns range from data privacy and surveillance to the potential misuse of information by state or non-state actors, as well as the unintended reinforcement of exclusionary or securitised approaches to migration and displacement. These risks underscore the need for critical scrutiny and robust ethical safeguards before integrating any new technological solutions into the search and identification of missing migrants. Those involved in developing AI tools used to search for and identify missing or deceased people in migration should adhere to the principles of trustworthy AI. These include inclusivity, explainability and reliability. Rigorous verification and validation of tools are recommended to build trust. Transparency is also essential in building trusted AI tools. This can be addressed through the concept of explainable AI, which requires making work in this field understandable and accessible to families and their communities, as described above. AI systems should not be “black boxes”. In addition, AI models must be trained on diverse, truly representative datasets to avoid perpetuating racism, discrimination, and social exclusion. New forms of collaboration are emerging among forensic scientists, families, academics, humanitarian engineers, international agencies, private actors and civil society organisations. These converging efforts to identify and locate missing migrants call for deeper conversations about responsibility, accountability, and the ethics of technological intervention. It is essential that civil society stakeholders, in collaboration with the research community and in dialogue with other stakeholders, apply their skills and expertise in the use of research and information-gathering technologies for identification, and contribute to making these processes increasingly understandable and participatory for families. Civil society stakeholders often serve as a bridge between families and other actors (including forensic scientists, researchers, international organisations, transnational institutions, national and local authorities and the private sector). Their role is essential both for providing concrete support to families and for legitimising their practical engagement with authorities to counter their monopoly over technology. These ‘new’ constellations of actors interested in developing new technologies to support the identification and search for missing or deceased persons in migration provide valuable insights that can enrich debates on humanitarian forensic action. I. INTRODUCTION Artificial Intelligence – encompassing a wide range of computational techniques, including machine learning, predictive analytics, and generative AI – and related new technologies appear to represent a new frontier. "An even more massive revolution than the internet revolution," says Mr Peter Honoiken (Finland, ALDE), Rapporteur of the Committee on Migration, Refugees and Displaced Persons at the Parliamentary Assembly of the Council of Europe. Within the council, Mr Honiken is promoting the Council of Europe Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law (2024), which claims to ensure the development of AI in line with human rights standards. Yet, the European Union (EU)'s massive investments in new artificial intelligence systems, along with the security-focused drift and its associated implications, have raised considerable concerns among civil society organisations (CSOs), non-governmental organisations (NGOs), migrant communities, and others. This is particularly evident on the issue of migration. As shown in recent EuroMed Rights reports, over the last three decades, border externalisation has been a core tenet of the European Union's strategy to address “migration flows”, especially those from the African continent (EuroMed Rights 2023; EuroMed Rights and State Watch 2023). The strategic geopolitical positions of the Middle Eastern and North African countries have positioned the region at the forefront of EU externalisation plans, making it a testing ground for new tactics and practices. In recent years, a new element has emerged in this externalisation strategy: the use of EU funds – including development aid – to outsource surveillance technologies that entrench political control over people on the move and the local population. The use of new technologies has been recognised as fundamental to the EU’s system of border control and migration management (EuroMed Rights & Statewatch 2023). Over the past three decades, the EU has built an extensive technological infrastructure for border control and migration management, relying on surveillance systems, biometric databases, and information networks. These digital border technologies serve two primary purposes: to facilitate “seamless travel” for “bona fide” visitors (tourists and businesspeople) with the “right” passports and sufficient income (Molnar 2024) in exchange for increasing amounts of personal data used to assess security risks, and to detect, deter, and expel specific categories of migrants considered undesirable - including refugees - through tools such LIST OF ACRONYMS ALDE AI AMD CSNA CSO DDA DVI EAAF EU EUAA EURODAC FMMI FRONTEX GHRC GSMA ICRC IOM MDVI ML NATO NGO OSINT SOCMINT SOCINT WGEID UN Alliance of Liberals and Democrats for Europe group Artificial Intelligence Ante-Mortem Data Complex Social Network Analysis Civil Society Organisations Databased Disappearance Analysis Disaster Victim Identification Argentine Forensic Anthropology Team (Equipo Argentino de Antropología Forense) European Union European Union Agency for Asylum European Dactyloscopy Europol European Union Agency for Law Enforcement Cooperation Forensic Missing Migrant Initiative Frontex European Border and Coast Guard Agency Global Human Rights Clinic Global System for Mobile Communications Association International Committee of Red Cross International Organisation for Migration Migrant Disaster Victim Identification Machine Learning North Atlantic Treaty Organization Non-Governmental Organisations Open Source Intelligence Social Media Intelligence Social Open Source Intelligence Working Group on Enforced or Involuntary Disappearances EUROMED RIGHTS - A study on the use of new technologies in the identification and search for missing or deceased persons in migration 11 EUROMED RIGHTS - A study on the use of new technologies in the identification and search for missing or deceased persons in migration as passport e-gates, drones, sensors, social media monitoring, databases and biometric registration. Public funding for these techno-borders has grown sharply. In the 2021–2027 EU budget cycle, funds for border policies have risen by 94% compared to the 2014–2020 cycle, with significant increases for countries like Greece, France, Croatia, and Spain. Billions are also invested in research and development, supporting projects on automated border gates, AI-based risk assessment, predictive migration analytics, and drone swarms. Primary beneficiaries include research institutes, private companies, and even the North Atlantic Treaty Organisation (NATO) bodies, with the EU having provided more than €250 million to 49 projects seeking to develop border technologies (EuroMed Rights & Statewatch, July 2023, pp. 5-6). Recent legislation (such as the EURODAC Regulation, the Schengen Borders Code reform, and the AI Act) is likely to reinforce these systems further, potentially encouraging racial profiling and expanding intrusive biometric practices. EU policy documents show no intention to reduce reliance on such border technologies; instead, they aim to optimise and extend their use, often bypassing political debate by framing changes as necessary technical measures. There is a risk that, disguised as a humanitarian approach aimed at protecting the most vulnerable - refugees and vulnerable populations, as defined by the new European Union Agency for Asylum (EUAA) vulnerability package (2024) - these measures may conceal forms that intensify security control at borders. In this report, we focus on analysing the potential implications of a different aspect of new technology use in the specific context of border deaths: the search for missing migrants along migration routes and the identification of border death bodies (Cuttitta and Last, 2020). Here, the deployment of new technologies has significant consequences for the families of missing persons seeking answers from European authorities regarding the fate of their relatives, and ultimately, their search for justice: for every nameless body recovered at sea, there is a family left uncertain whether their loved one is alive or dead. In that respect, the growing number of conferences, workshops, and expert meetings on the use of artificial intelligence (AI) in the search for missing migrants constitutes a relatively recent yet noteworthy development. This trend underscores a growing interest among diverse stakeholders in exploring technological innovations as potential tools to support efforts to trace and identify missing persons. One prominent example is the organisation of a recent symposium titled “New Technologies and the Missing, including Victims of Enforced Disappearance: Developing Opportunities and Navigating Risks,” held in Geneva on June 24–25, 2025. The event was co-organised by the International Committee of Red Cross (ICRC) Central Tracing Agency, the ICRC Global Cyber Hub, the UN Working Group on Enforced or Involuntary Disappearances (WGEID UN), Luxembourg Aid & Development, and the Swiss Federal Department of Foreign Affairs, in collaboration with the Global Alliance for the Missing. Bringing together over one hundred participants - including representatives of state authorities, search practitioners, investigative institutions, academic scholars, family associations, prosecutors, NGOs, and the private sector - the symposium featured several sessions specifically dedicated to AI, machine learning, and predictive analytics. A few weeks later, the Migrant Disaster Victim Identification (MDVI) Workshop “AI Fundamentals and Applications in the Search and Identification of Missing Persons” was organised on 9–11 July 2025, hosted at the Research Centre in Information and Communication Technology at the Universida de La Coruña (UDC) in Spain. This three-day workshop examined technological applications ranging from the use of social networks in search efforts to natural language processing, image acquisition and processing, machine-learningbased image comparison and registration, and the principles of trustworthy AI. Taken together, these initiatives highlight how the specific issue of missing migrants in the case of dangerous journeys is increasingly situated within the broader field of missing persons, including 13 those who have disappeared as a result of armed conflict, violence, or natural disasters, as well as within the broader scope of the development and analysis of AI. Given the scale of fatalities among those attempting to reach Europe - estimated at approximately 30,000 deaths in the Mediterranean since 2014 (UNHCR; IOM) - migrant deaths are increasingly framed as events requiring a Disaster Victim Identification (DVI) response. DVI protocols as standardised processes of identifying victims of mass-casualty incidents (such as natural disasters, aviation accidents, or industrial failures that involve more than 5 casualties) are therefore being reviewed in the context of dangerous migration journeys, taking into account some of the challenges involved that are linked to post-modern conditions, as well as the availability of antemortem information. In this context, the MDVI COST Action programme, established in 2023, was launched to convene stakeholders across Europe in addressing the growing humanitarian crisis of unidentified deceased migrants. Within the MDVI COST Action framework, more than 198 members representing hundreds of institutions across 35 countries collaborate to develop, standardise, and validate international processes, resources, and methodologies. These include integrating innovative tools such as craniofacial identification techniques, drone technologies, and AI technologies. As we will see in this report, some of these technologies have already been successfully adopted in the search of missing migrants and identification of border deaths. However, whether they are official DVI procedures or the cutting-edge methods specific to MDVI, it should be clarified that the state regulates this type of intervention. Indeed, it is the public prosecutors and local government authorities who request the involvement of forensic experts and specialised protocols. In the absence of a mandate to identify disaster or border victims, these protocols are rarely used. Currently, only a minority of missing migrants are identified, and numerous factors contribute to this ongoing human tragedy. Up to now, the principal challenges include the absence of national and international regulations mandating the identification of bodies, a lack of political will to implement identification commitments, and a lack of protocols or funding to support migrant identification processes. Additional obstacles are logistical in nature, such as limited communication between countries (countries of origin, transit states, and destinations) and inadequate coordination among key stakeholders (including family members, migrant communities, advocacy, forensic experts, government agencies and humanitarian organizations). Compounding these issues are infrastructural deficiencies, including the recovery and dignified storage of bodies, access to appropriate facilities for analysis, and proper burial arrangements, all of which exacerbate the already fragmented management of this phenomenon. In this context, the digital revolution has dramatically expanded the volume of available information - particularly through social media networks, geospatial imagery, and digital archives - thereby amplifying the perceived relevance of AI-driven approaches for different stakeholders in the field. Machine learning, generative AI and predictive analytics are well-positioned to process large datasets (combining old and new data), enabling the extraction of useful information, the identification of patterns, and the recognition of connections across cases. Efficiency and timeliness are frequently invoked as justifications for integrating these new technologies into search practices (ICRC Communication Report, 2025). This approach is also the one taken by the Parliamentary Assembly of the Council of Europe. On October 3rd 2025, Resolution 2628 (2025) on Artificial Intelligence and Migration was adopted, referring explicitly to AI on the issue of missing migrants: “The Assembly expresses once more its deep concern and sorrow over the deaths at sea of migrants, refugees and asylum seekers. It calls on member States to employ AI technologies to enhance search and rescue capabilities, and to uphold the dignity of deceased individuals, in accordance with the principles laid down in its Resolution 2569 (2024) “Missing migrants, refugees and asylum seekers – A call to clarify their fate” and Resolution 2612 (2025) “Saving the lives of migrants at sea and protecting their human rights”. EUROMED RIGHTS - A study on the use of new technologies in the identification and search for missing or deceased persons in migration However, the use of AI in the search for missing migrants and in the identification of the bodies of deceased migrants presents significant challenges. It raises crucial legal and ethical concerns for the families of the missing, search operators, and others whose personal information may be exposed without consent. These risks include the deployment of poorly validated models, algorithmic bias, privacy breaches, potential misuse for criminal purposes, and broader issues related to data protection and privacy. This report aims to trace the development of these new technologies in the search for missing migrants and in the identification of the bodies of deceased migrants. It examines the debates their use has generated regarding associated risks among the different actors involved in this field (academics, NGOs, IOs, private actors, CSOs, etc.). The findings in this report are based on a combination of desk-based research and qualitative interviews. We reviewed academic literature, policy documents, reports from international organisations, and media sources on topics such as forensic science, migration, and emerging identification technologies, with a focus on the Mediterranean but drawing on global examples where relevant. This was complemented by semi-structured interviews with 35 stakeholders, including forensic experts, institutional representatives, civil society organisations, legal specialists, and family or community members, selected through existing networks and snowball sampling. By synthesising insights from both literature and interviews, we map the use of new technologies in identifying missing or deceased migrants, and critically assess their potential benefits and risks, particularly for families and CSOs. That said, it is essential to note from the outset that, despite significant attention to AI's potential, these technologies have not yet achieved widespread adoption. Their application remains fragmented, primarily because experimentation is confined to a limited number of actors, with minimal collaboration among stakeholders. 15 II. FORENSIC APPROACHES TO THE SEARCH FOR MISSING MIGRANTS AND THE IDENTIFICATION OF DECEASED MIGRANTS To explore this element further, beyond analysing new technologies on the one hand and the phenomenon of deceased and missing migrants along migration routes made illegal by EU policies on the other, it is necessary to examine the role of forensic experts who mediate between the demands of families and institutions. Since the mid-1980s, forensic scientists have played a crucial role in the international response to mass violence, contributing evidence to war crimes tribunals and identifying bodies to end the torturous uncertainty of loved ones. Forensic intervention has also taken on a particular dimension along European borders, especially since the intensification of public attention to border crossings and deaths in 2015. The growing number of deceased or missing migrants, and unidentified bodies along the Mediterranean route has attracted an increasing number of forensic scientists, international and non-governmental organizations, civil society actors, and various research interventions to investigate the European humanitarian crisis and its victims (Kovras and Robins 2016; Ben Attia et al., 2016; Squire et al., 2017). These actors, engaging in what Moon (2022) has described as extraordinary “death-work,” driven by their technical expertise and the desire to apply their knowledge to the context of Mediterranean migrant identification, have either modified existing procedures to adapt to this new context or designed and implemented new technologies aimed at identification, seeking to address the challenges that the identification of deceased migrants presents (Baraybar et al., 2020). But what do we mean when we talk about forensic intervention? Different genealogies help us understand this phenomenon. The emergence of humanitarian forensic action is commonly traced to Argentina in 1984, when a delegation of forensic scientists supported the newly established truth commission, which established the National Commission on the Disappearance of Persons to investigate the fate of thousands of disappeared persons under military rule (1976–1983). At the request of the Mothers of the Plaza de Mayo, this mission led to the creation of the Argentine Forensic Anthropology Team (EAAF), the first independent group dedicated to applying forensic anthropology to cases of political violence (Rosemblatt 2015). From this starting point, humanitarian forensic practice developed along different trajectories. The EAAF’s work combined a commitment to families EUROMED RIGHTS - A study on the use of new technologies in the identification and search for missing or deceased persons in migration with an orientation toward accountability and justice. This “Latin American model” embraced what Mercedes Doretti, co-founder and full-time member of the EAAF team, called the “grey spaces and endless negotiations” of transitional contexts, pursuing truth, justice, reparation, and prevention together (Doretti & Burrell 2007). Another branch of the genealogy of forensic intervention, to use Adam Rosenblatt’s (2019) expression, is represented by the work of the ICRC, which in the early 2000s institutionalised its own forensic program around the category of “missing persons.” The ICRC drew a sharper line between humanitarian action and judicial accountability. Tidball-Binz, a former coordinator of the ICRC's forensic unit and EAAF founding member, framed the work in terms of neutrality and the “right to know” the fate of the missing (Cordner and Tidball-Binz, 2017). The ICRC emphasised identification and repatriation of remains while avoiding direct involvement in prosecutions. This divergence between the EAAF and ICRC reveals a fundamental tension in humanitarian forensics: whether the task is primarily to end the torturous uncertainty faced by the families of the disappeared or to contribute to legal accountability. If these interventions have traditionally taken shape in contexts of enforced disappearances, genocide, and mass disasters, for this discussion it is particularly relevant to consider the work of anthropologists at the U.S.-Mexico border as an example of forensic humanitarian action that transcends traditional frameworks and focuses instead on identifying border deaths, as the work of Anderson (2008), Reineke (2016 & 2022), Soler and Beatrice (2017) have outlined. These researchers describe migrant deaths as the result of U.S. “prevention through deterrence” policies that push migrants into dangerous desert regions. This approach expands forensic humanitarian action beyond traditional conceptions of humanitarian emergencies, engaging directly with global wealth distribution and its impact on who lives and who dies. Building on and, in some cases, running parallel to these approaches, the term counter-forensics has also emerged to describe an orientation that employs technical evidence –typically utilised by experts in judicial proceedings – to establish alternative narratives that challenge dominant accounts. Within the counter-forensic framework, individuals and CSO repurpose forensic tools and methodologies to examine state practices and generate new forms of truth and political engagement (e.g., Pezzani and Heller 2013; Keenan and Weizman 2012; Weizman 2014). An example of this approach is the work of the Border Forensics project,1 which operates within the broader Forensic Oceanography initiative, itself established under the auspices of Forensic Architecture. These projects employ opensource data and tools, digital modelling, and immersive technologies to investigate border violence and foster mobility justice. The results are impactful, as they deploy interactive maps, videos and other visualisations as evidence of border violence and migrant deaths. It is also important to underline that the birth of international forensic action and the proliferation of these different approaches coincide with what scholars have termed the “forensic turn” in social sciences - the normalisation of forensic science, exhumation, and burial practices in contexts of atrocities and mass deaths. This turn has several defining features that go beyond the purpose of this report (for more details on this, see Dreyfus and Anstett, 2014; Weizman, 2017; Dziuban, 2020). Here, however, we would like to emphasise a few salient aspects for our analysis. First, the forensic turn elevates material traces - the body, the bone, the archive of remains—as the privileged site of truth. Second, it grants forensic experts an “interpretive monopoly,” positioning them as the sole legitimate mediators between mute material evidence and public meaning (Crossland, 2018). Third, it displaces the political and collective dimensions once embodied in courtroom testimony: the witness as political subject and the trial as forum for shared truth-making give way to 1 https://www.borderforensics.org 17 expert authority and technoscientific objectivity (Mazzucchelli, 2017). The expansion of forums for presenting forensic evidence - ranging from truth commissions to art installations, museums, and humanitarian reports - has been described under the notion of “forensic aesthetics”: “forensics is not only a matter of science but also of presentation: the making public of evidence across a multiplicity of forums” (Weizman, 2017). While this multiplication of arenas creates new opportunities to disseminate knowledge about violence, it also privileges expert interpretation over lived testimony. Survivors and witnesses risk being displaced by the authority of technical expertise, as material traces are framed as the ultimate bearers of truth. This shift carries the danger of depoliticizing violence and suffering, reconfiguring them as objects of scientific or aesthetic interpretation rather than matters of collective accountability. In this sense, the humanitarian forensic project not only broadened the reach of forensic science beyond domestic criminal law but also reconfigured the very politics of truth. An “excessive focus on forensic identification” (Martinez-Garcia et al., 2024) might overlook the broader context of missing migrants and the role and agency of families in the search. The investigative work that families undertake to find their missing loved ones is not insignificant. Research has found that families often initiate or drive the search process, despite many obstacles (Okyere and Kondeh, 2021). It is therefore essential to recognise this agency rather than viewing the families as passive victims or merely sources of antemortem data, and also to acknowledge the stance families are taking against the inaction of the state to identify their loved ones (Reineke, 2022). Perhaps the ultimate example of families taking control of their search and influencing the practice of others, including forensic scientists, comes from the Latin American context. Referred to variously as forensic civism, citizenship, or citizen-led forensics (Schwarz Marin and Cruz Santiago, 2016; Cruz Santiago, 2020), this approach involves community-led investigations in response to state authorities' inaction. To date, the examples of family-led actions in Mexico, Colombia and along the USA-Mexico border show the importance of including families of missing migrants in the research for identification. There is, therefore, no singular forensic paradigm but a plurality of approaches grounded in distinct local contexts. Each is constituted through specific configurations of humanitarian, political, and legal logics, which variously expand or circumscribe the possibilities for family and citizen participation. It is against this backdrop that the emerging deployment of AI-based forensic technologies - designed to locate and identify migrants who have died or gone missing in the Mediterranean - should be understood. 1. The case of missing migrants in the Mediterranean represents a sui generis scenario The case of missing migrants in the Mediterranean represents a sui generis scenario, as explained by José Pablo Baraybar do Carmo, an ICRC Peruvian forensic anthropologist who has been instrumental in much of the evolution of forensic science outlined above. First and foremost, the fundamental trace essential to forensic work is often absent: the body itself. According to the ICRC, only approximately 15% of the bodies of migrants who die at sea are recovered by authorities (2022). Additionally, there is uncertainty about the location of disappearance – whether it occurred at sea or before departure – given the numerous physical and material borders migrants must traverse, including desert crossings and detention facilities along the EU's externalised borders. The condition of recovery remains further complicated by forensic work. Bodies retrieved from the sea, for instance, often undergo saponification – the transformation of body tissues into a soap-like substance due to fat decomposition – which impedes DNA extraction and facial recognition. To these challenges must be added issues related to the storage of remains, as well as the structural problems previously highlighted, including the absence of legal frameworks mandating identification, EUROMED RIGHTS - A study on the use of new technologies in the identification and search for missing or deceased persons in migration Andrea Garcia, from IOM, was more cautious in her assessment of the DVI response to Pylos: “I don't know the percentage, but it's like a really high percentage of the people that were identified from that specific shipwreck. But I have a problem with this being used as an example of a good practice or something like that because it was very successful to identify the bodies that had been recovered. But the large majority of the bodies were not recovered.” 4. DNA “There is this understanding that genetics will solve everything…” Mercedes Salado Puerto, EAAF. Since its development in the 1980’s DNA profiling has become a standard forensic technique in domestic criminal justice settings and also international identification efforts, to the point that it has gained an almost “transcendent evidential quality” and is often given “epistemic priority” (Lynch et al., 2009, p.xii), to the point that DNA profiling has become synonymous with forensic identification (Bennett, 2014). It is one of Interpol’s primary identifiers, and some efforts to identify the dead, such as those following the conflict in the Balkans in the 1990s, have been classified as DNA-led. It must be noted, however, that the use of DNA testing is not apolitical. The resources and funding for DNA may be provided, not only for humanitarian reasons, but rather as a “result of a desire for the political and social capital that this highly prestigious technological device offers, to individuals, organisations, and governments involved in its deployment” (Bennett, 2014, p.240). As Mercedes Salado Puerto of EAAF commented: “The state is focusing on genetics, 100%”, agreeing with a member of the Border Forensics team who referred to the “Mystification of the power of DNA”. Given its long-established use, DNA profiling cannot be considered a new technology and therefore falls outside the scope of this report. Moreover, as discussed above, there are specific limitations to the use of DNA evidence in the context of migration, such as the lack of ante-mortem samples and the severely degraded condition of many bodies. However, there are some recent developments in the field which hold promise or are being proposed to aid in the identification of deceased migrants. Some recent developments in the field of forensic genetics, which could be classified as “new technologies” for this report, are outlined below: 5. Rapid DNA Standard laboratory procedures typically require 8–10 hours to generate a DNA profile. However, forensic genetic laboratories rarely process single samples in isolation; given the workflow and sample volume, generating a DNA profile can take several days. Since 2015, improvements in technology and microfluidics have enabled rapid DNA processing, allowing a DNA profile to be generated in as little as 2 hours (Romsos & Vallone, 2015). In addition, the instruments used to create these rapid DNA profiles are generally portable and can be deployed at crime scenes, in custody suites, and even as part of a DVI response (Bowman et al., 2022; Forensic Science Regulator, 2025; Hares et al., 2020). There are also reports of rapid DNA being used for family reunification in relation to the ongoing conflict in Ukraine (UNDP, 2025). Another reported advantage of rapid DNA technologies is that they are designed for non-technical users, and the training for operators is minimal (Kaplan et al., 2023). While some hold concerns that 31 DNA samples should only be processed by accredited forensic laboratories and the results interpreted by experienced forensic scientists (Dolan, 2019), the potential for non-expert users does allow for non-governmental groups (e.g., human rights or humanitarian organisations) to carry out their own DNA tests in sensitive contexts, such as that of missing migrants (Madden & Katsanis, 2021). Kaplan et al. (2023) report successful identifications of previously unidentified human remains, believed to be migrants in the USA, from skeletonised material, using rapid DNA. Although a 2023 study by Chong et al., found conventional methods were more successful than rapid methods for DNA typing of compromised bones. Genetic information is increasingly being used at the USA border. In their study, Madden et al. (2021) found that the potential pitfalls of using rapid DNA technology in this context outnumber the utilities. One particular issue is the geneticization of the concept of family and the authors caution that family relationships do not correspond to genetic relationships in the same way across languages and cultures. Nonetheless, some are calling for the roll-out of rapid DNA at European borders (Schroeder, 2025), although to our knowledge, this is not happening in Europe to date. One research participant, Luis Fondebrider, Independent Forensic Consultant, former EAAF/ ICRC, used rapid DNA as example of a new technology being sold or viewed as a “magic solution” and urged caution: “ Until today it is in use, in very recent cases, but with bones and teeth it doesn't work like that. But they are selling the rapid DNA like the magic tool to solve every kind of gain in Gaza, in Ukraine. Whatever country it is, there are companies selling this. There are people buying this because this is a magic solution. But when you go to the detail you see it's not a magic solution.” 6. i-Familia Despite the issue of potentially limited availability of antemortem samples for comparison to a DNA profile generated from an unidentified body, DNA does offer one advantage over the other primary identifiers: you do not need a direct reference sample for comparison. Because DNA is inherited, indirect, or kinship comparisons between family members and a questioned sample are possible. This process generally involves complex calculations and relies on population-specific data. This presents challenges in the migration context, where you might not know the population an individual belongs to or have access to a data set for that population. While DNA-matching databases are not new technology, in 2022 Interpol launched the i-Familia database, a global database for identifying missing persons based on international DNA kinship matching. This database could be considered a new technology because, to overcome the challenges above (regarding population genetics), they had to develop an innovative methodology that involved the DNA-matching software BONAPARTE, worldwide allele2 frequencies, and tailored statistical thresholds (Laurent et al., 2022). From the perspective of scientific validation and transparency, it is encouraging to see this methodology published in a peer-reviewed scientific journal. While the technological advancements related to i-Familia are promising, there are concerns that it is a tool developed and held by Interpol, a law enforcement agency. Amankwaa et al., (2025) highlight ethical, legal, societal and privacy concerns, including function creep. The i-Familia brochure says that families of missing persons and missing persons associations who want to contribute DNA samples to i-Familia should “contact your national police” (Interpol, 2022). However, it is documented in 2 Allele = the version of a DNA marker which an individual has (inherited from parents). The individual alleles combine to make a DNA profile. EUROMED RIGHTS - A study on the use of new technologies in the identification and search for missing or deceased persons in migration the literature that families of missing migrants may experience fear or anxiety about approaching state authorities in their search and may hold a mistrust of the police (Reineke, 2022). Jose Pablo Baraybar, from the ICRC, also expressed concerns about framing DNA collection primarily through law enforcement perspectives, noting the potential mismatch when agencies such as Interpol are involved in humanitarian contexts and seek access to DNA databases. Andrea Garcia, from IOM, shared similar concerns about Interpol, as a police organisation, managing the DNA database most of use to families of missing migrants. Although she did also acknowledge the potential for successful matching through formal mechanisms, with international reach: “It could be effective, but the problem is that the police need to be involved and, as we all know, people searching for missing migrants are most likely not going to go to the police. So there's something that is already there, but it doesn't serve the purpose of missing migrants specifically.” Florian von König (ICRC) highlighted i-Familia as the only existing platform for comparing international DNA profiles in missing migrant cases, noting that it is designed as a humanitarian database with restricted access. Encouraging its use, he emphasised that it remains the only free tool available for international DNA matching. Unlike other Interpol databases, i-Familia allows only a small number of Interpol personnel to conduct DNA kinship matching and notify the relevant states if a match is found. Yet, despite its potential, the platform faces significant challenges, and among them the lack of profiles from families in most countries of origin, particularly from Africa: “most current migrationrelated profiles appear to be coming from families based in Europe. The fact that the system is only accessible through enforcement actors is another potential hurdle to families”. 7. DNA & AI There are examples in the literature of AI applied to forensic genetics for human identification, including for predicting an individual’s biogeographic ancestry (e.g., Barash et al., 2024; Marisco & Amigo, 2025). Tania Delabarde from Paris Cité University cited this as a promising use of new technology, which she had experienced in her forensic casework. Access to genetic genealogy techniques, including the inference of phenotypic traits, may help estimate the likely geographic origin of unidentified individuals. Such genetic information can help narrow the provenance of unknown biological samples, including human remains. However, the analysis generally only indicates a broad geographic region (Phillips, 2015) and it is essential that genetic data is not conflated with social constructs such as race or ethnicity (Marisco & Amigo, 2025). Another participant, Jan Bikker, from the Forensic Missing Migrant Initiative (FMMI), discussed a further potential intersection between DNA analysis and AI that could be useful for identifying missing migrants. He noted that, because parents or close relatives are often unreachable – frequently due to conflict, displacement, death, or reluctance to collaborate – investigators must usually rely on distant relatives. Emerging AI-driven developments in genetic analysis, he suggested, may facilitate the use of mitochondrial DNA – the small, circular genetic material found in mitochondria, inherited almost exclusively from the mother, therefore helpful in tracing maternal lineage and identifying distant biological relationships – or other markers to conduct predictive analyses that could improve identification outcomes when close-kin reference samples are unavailable. However, Bikker also emphasised the significant challenge of communicating such results to families. Even in straightforward cases involving close biological relationships, families often struggle to 33 understand how identifications are made; introducing AI-assisted inferences would further complicate this process and undermine trust. As he put it, if explaining a direct parent–child match is already difficult, explaining an AI-derived match will be “very, very tricky.” Here Jan raises another critical point in relation to DNA analysis, which is how it is explained to families in a way that they can comprehend and then decide whether they trust and accept that result. The involvement of AI would further complicate this. 8. Fingerprints Fingerprints are another primary identifier used by local police and Interpol. They are routinely used for identification in criminal casework and for identifying the deceased. However, recent technological developments in fingerprint collection hardware and also comparison software could see applications in the forensic identification of missing migrants. For example, improvements in fingerprint-scanner technology enable more explicit fingerprint images (Tamisier et al., 2019), while new matching algorithms offer greater accuracy (Tom et al., 2022). The use of fingerprints for identification following the 2023 Pylos shipwreck was cited above, particularly in relation to Pakistani nationals. Andrea Garcia (IOM) noted that during a recent meeting of national focal points on missing migrants within the Rabat Process, states began exploring which data sources could support future identification efforts. Among the options discussed was the use of biometric data collected at border crossings. Although many migrants who die during their journey may have travelled irregularly, this does not preclude the possibility that they previously crossed borders legally – particularly in regions such as West and Central Africa, where the Economic Community of West African States free-movement arrangements apply. In such cases, their fingerprints or other biometric information may already exist in national databases. Participants also underscored that the ideal long-term solution would be a transnational database combining antemortem data (including border biometrics) with any post-mortem information recovered from unidentified remains. Given the rapid expansion of biometric border systems in the last decade, Andrea believes that these datasets could become an essential resource for improving future identifications. Jan Bikker (FMMI), also mentioned the fingerprint data collected during border surveillance, although access to this data may be an issue, and how AI could be used to enhance fingerprints: “Biometrics, fingerprints, for example, are still very useful for us, so I think that is still a very good tool. And of course AI techniques, now they can actually do a lot of good work for decomposition [...]but the problem is that, when it comes to border security, it's very difficult to get any data.” As Jan emphasised, the use of fingerprints raises fundamentally ethical and political questions about who collects this data and for what ends. The ethics of repurposing biometric data hinge on principles such as informed consent, transparency, privacy, and accountability—issues examined later in the report. 9. Facial Identification Given that many countries hold large numbers of unidentified human remains—suspected to belong to migrants—and face significant barriers to identification due to the limitations of established forensic techniques in this context, it is increasingly evident that new approaches are required, EUROMED RIGHTS - A study on the use of new technologies in the identification and search for missing or deceased persons in migration extending beyond Interpol’s primary identifiers. Emerging identification techniques developed through investigations of criminal cases and other human identification efforts could be applied to missing migrant cases. Similarly, “new technologies” based on AI are being proposed as solutions in this space (Panacea, 2023). AI applies to various fields within forensic anthropology, or skeletonbased forensic identification, such as biological profile estimation, trauma description, and visual comparison (Mesejo et al., 2020). Forensic facial comparison, which generally involves comparing the face of an individual from a photo or video still to a reference image of the suspected identity (Wilkinson et al., 2024), is not a primary identifier, nor would it be considered a new technology. However, new developments are being tested in the migrant identification space, and this technology was mentioned during the research interviews. Given people’s online profiles, including photographs, selfie photographs from social media could be considered a source of ante-mortem data for post-mortem identification (Miranda et al., 2016; Caplova et al., 2017). This methodology was applied following the Lampedusa shipwrecks of 2013 (Olivieri et al., 2018), although with a low identification rate. Research in the field is ongoing, and a new multi-million-euro project, funded by the European Research Council (ERC), is related explicitly to naming dead migrants in Europe. The FIND ME3 Project (Forensic anthropological Identification to Name the Dead Migrants in Europe) led by Cristina Cattaneo at the Univesity of Milan (2025) aims to “advance the field of forensic identification by focusing on the anthropological analysis of morphological features of the face” (University of Milan, 2025). Historically, the process was manual, but newer tools such as the craniofacial superimposition modules of Skeleton-ID (Panacea, 2025) offer patented skull-face overlay algorithms that automatically compare skulls to images in seconds. 3 https://scibis.unimi.it/en/research/funded-projects/forensic-anthropological-identification-name-dead-migrants-europe-find-me 35 “Many things are still in development. And like I said, when you get big grants, there's only very few in Europe that can actually get big grants and do potential research projects like AI and face recognition. And there's a lot of really interesting technology in terms of face recognition, facial reconstruction even, or in the matching as well [...]even predictive modelling, like how decomposition affects the face and then refers to it to match it to people [...]but I think it's very much new and if you want to develop something, it's going to take four years, five years....” One example of these emerging technologies is Skeleton·ID, a software platform developed within the Panacea framework and currently accessible online. Skeleton·ID is the only tool that supports skeleton-based identification through a suite of physical anthropology techniques, including automated craniofacial superimposition, which overlays skulls with ante-mortem photographs to assist in identity assessment (Panacea, 2025). Another specific new technology tool, related to facial identification, which was discussed during our data collection interviews is an app, imbox, being developed by scientists at INSA Toulouse, in collaboration with the ICRCg. Jules Ripoll, who is the PhD student working on this project, explained what it involves: "I work with AI models to try and reconstruct faces for people who died of severe injuries. I work with all kinds of brutal deaths. Basically, I use machine learning AI tools that know very well what a human face is. So we leverage this knowledge to rebuild the parts of the faces that are broken. It's in a pretty automatic process, also efficient, smooth. In this way, we don't have to spend a lot of time working manually on the image, and so we can provide nicer images to be then identified." The researchers also hope to make the app open-source and available to any organisation that wants to use it. Currently, the application has been used in the field only sporadically. For example, during an interview with MemMed, an activist explained that they did not use the INSA app directly but instead sent a photo to the developers, who returned an enhanced image. This tool proved essential in identifying a migrant who had died at sea: the face in the original photo was unrecognisable. At the same time, the reconstructed image allowed a relative to examine it and provide additional information for identification. It is important to note that this identification occurred outside official forensic procedures. The process was entirely handled by civil society actors – MemMed, as a mediating association between families and authorities – using unofficial and not yet widely available tools. Reports are emerging from the ongoing conflict in Ukraine of the controversial facial-recognition technology Clearview AI being used to identify the dead (Bhuiyan, 2022). However, concerns surround the software as it is unclear how effective it is at recognising the dead, the technology may be racially biased, and it is being contested in various countries due to data and privacy concerns (ibid). In addition, as facial recognition solutions emerge and are beginning to be applied to the field of identifying deceased migrants, there is a focus on biological identity, which could overlook the important concept of an individual’s social identity and the involvement of families in the process. This was highlighted by the member of the Border Forensics team we interviewed: “I am a little bit scared about this idea of facial recognition, for example. Because the problem I think it would be if you implement the capacity to identify a body from the same body… I mean if you capture the image from the person online or on the boat and then you check it and you match it with the body. You can do an identification of the body without the biography of the person, so you didn't define the person.” Figure 6. Panacea - https://www.youtube.com/watch?v=YA69Cva1xFg Jan Bikker (FMMI) emphasised that while these new technological approaches may have promise, in many respects we are still in the developmental stages, far from having operational tools: EUROMED RIGHTS - A study on the use of new technologies in the identification and search for missing or deceased persons in migration 10. Age Estimation Similarly, AI models have been designed for automated age estimation and were mentioned by participants during the data collection interviews. These models are thought to “demonstrate significant promise” (Khanagar et al., 2024). In fact, AI models have been shown to outperform the classical (human) methods applied for age estimation (ibid.). These AI tools enable fast multiple comparisons and efficient filtering, reducing the number of potential candidates in a database very quickly. Additionally, some argue that AI tools can be trained to remove human bias (Mesejo et al., 2020). However, researchers caution that studies with larger sample sizes are required and acknowledge that designing and training a deep neural network is an expensive, time-consuming, and complex process. The precision of models depends on the quality of training data and training procedures (Khanagar et al., 2024). These tools are unlikely to replace forensic practitioners, but could be a valuable partner to (human) professionals. In a space where research funding can be challenging to secure, it is interesting to note that there are age estimation tool research projects funded by the EU, such as ‘Unaccompanied Minors Automatic Forensic AgeEstimation’ (UMAFAE, 2022-2024)4 and ‘Multimodal abd Multifactorial Image-based Age Estimation’ (M2IbAE, 2025-2027), 5 coordinated by Panacea Cooperative Research. 6 These initiatives explore the importance of multifactorial and multimodal approaches for advancing forensic age estimation. Age estimation can contribute to the forensic identification of an individual, but there are concerns it can also be used as a “technology of control” as those judged under 18 may have greater protection in the asylum system of a country (Dahler, 2020). Óscar Ibáñez Panizo from Panacea Cooperative Research explained to us: “In the European Union… They are more interested, I think, in border control, but not identification at the border […] And indeed we get some funding for example for age estimation of minors, legal age estimation. Because, well, it could be a technology that can be eventually applied in the world, although what it is more is to try to help young people to get their rights.” This is an example of a new AI technology that shows promise at the scientific level. But its use is contested as to whether it will actually “help young people to get their rights” or be used as a “technology of control” and this will be determined by who is using it and to what end. 11. Data & Databases The harmonisation of data collection through a standardised dataset has been a topic of discussion among field actors, often viewed as a means to enhance responses to the issue of missing migrants. In fact, a large amount of data on missing migrant cases may already be available across different repositories. Over the last few years, specific databases have been developed specifically to count border-related deaths. Among these, we highlight two examples. The first is the “Deaths at the Borders Database”,7 developed at Vrije Universiteit Amsterdam as part of a research project led by Tamara Last and directed 4 https://cordis.europa.eu/project/id/101026482/reporting 5 https://cordis.europa.eu/project/id/101209534 6 https://panacea-coop.com/ 7 http://borderdeaths.org/ 37 by Thomas Spijkerboer (Last et al., 2017). This database constitutes the first systematic collection of official, state-produced evidence on people who died while attempting to reach southern EU countries from the Balkans, the Middle East, and North and West Africa, and whose bodies were recovered in, or transported to, Europe. The database covers the period from 1990 to 2013, while the ICRC later updated it with data spanning 2014 to 2022. Next to it, we can find the IOM database as part of the "Missing Migrants Project",8 which has documented over 72,000 deaths and disappearances on global migration routes since 2014, using non-state-produced evidence. However, this database falls short of providing the necessary information for families to identify their loved ones, and, most importantly, it is not directly accessible to the migrants it documents. There are numerous issues with the data related to missing migrants, which are impeding identification efforts. Data may be held in disparate locations by organisations that do not trust each other and therefore do not want to share information. Some have noted this fragmentation, such as Cattaneo et al. (2022), and the ICRC has begun the vital work of standardising data through its core dataset for the search for missing migrants (ICRC, 2021). In that respect, ICRC Missing Persons Project9 was envisioned as an entity capable of collecting and harmonising data from various sources, while ensuring the protection of this sensitive information amid tight border security measures. Unfortunately, despite several attempts, this project has ultimately been abandoned, ultimately for lack of funding. Tania Delabarde, from Paris Cité University, explained the importance of databases to connect information and facilitate identifications: “So for me, what is very frustrating is the lack of access to other agencies that are looking for someone. I understand all the legal limitations, all the data protection, I completely understand all that. But it means we have people somewhere, someone is looking for them and there is no link.” However, it is the links between data that allow identification. Andrea Garcia, from IOM, emphasised this while suggesting it may be possible with existing technologies: “There should be more exploitation, for lack of a better word, of the data that's already collected or that's already available like biometric data. Any kind of databases that countries have on their citizens, etc. It's not so much about coming up with a new technology, but more about how we can use what's already there.” But, as Florian von König, (ICRC), explains there could be ramifications of this approach: “It's a double-edged sword. It's tempting because you have these ever growing repositories of different biometric data etc., that's obtained from migrants or about migrants and some of which could be useful for identification and search efforts. But if we try to use this information for humanitarian identification purposes, will that reinforce the sense that we should get even more of it? Because there's also humanitarian benefits and that can be used in all kinds of problematic ways to say, we're collecting whatever iris scans, fingerprints because it's also a humanitarian measure. So it's a slippery slope.” Lanzarone et al., 2025 describe some of the different biometric databases held around the world, such as Pakistan’s national fingerprint database, and how they can provide crucial identification leads. These biometric databases may be held for administrative, electoral, immigration, criminal or 8 https://missingmigrants.iom.int/ 9 https://www.icrc.org/en/publication/4375-missing-persons-project EUROMED RIGHTS - A study on the use of new technologies in the identification and search for missing or deceased persons in migration other purposes. The authors claim that access to these kinds of databases could help identification efforts, but caution that data protection is an important consideration. O’Regan (2025) highlights the importance of data protection legislation for protecting migrants. It is almost inevitable that their biometric data will be collected during their journey. Still, digital privacy may be a low priority for them amid the multitude of other challenges they face. In their 2023 report on New Technologies, the WGEID UN identified that “technologies currently employed at borders for migration control [...] could conversely be applied to facilitate the search for disappeared people”. They conclude that it may be worth exploring this aspect of these technologies, provided fundamental human rights are respected (WGEID, 2023). In their input to the WGEID thematic study on new technologies, the Border Violence Monitoring Network reiterated that new technologies have the potential to be used to investigate and even prevent disappearances. Still, they also raise critical human rights concerns (BVMN, 2023). Databases themselves may not be classed as new technologies, but many that are used to manage data from missing persons cases, and that are either currently used or have the potential to be used in missing migrant cases, were mentioned during our data collection. These included NamUS, the ICRC Resolve platform, Plass Data, Bonaparte, M-Fysis, ICMP’s iDMS and i-Familia (Interpol). These are traditional databasing software solutions. But as Óscar Ibáñez Panizo, from Panacea Cooperative, explained: “Those are mainly databases with limited search and comparison capabilities.” Some offer more novel functionalities than others, such as the NamUS database in the USA,10 which provides limited public access. Tania Delabarde, from Paris Cite University, explained how her attitude toward this type of “open” database changed over time: “When I started my career I was really questioning myself regarding some kind of database like NamUS in the USA, I was saying “Oh my God, that's horrible. You have all these photos of dead people and then people can access it”, but now I have to say, after all these years, I think that's the best solution. I mean, I'm really now involved and say we have to let the relatives access the information.” If the power of databases is combined with new technologies such as AI and machine learning, there may be even more potential for vital connections to be made as Anjli Parrin, from the Global Human Rights Clinic at the University of Chicago Law School, discussed: “When you have mass amounts of information the problem is actually organisation and structure rather than quantity. I think there are really positive uses of technology in the analysis, sifting through, and creating of good databases for understanding and making sense of data. And then finally technology is useful for comparing different data sources”. This kind of approach has been described in the literature as well. Reyes et al., (2025) outline a new type of applied research, which they are calling databased disappearance analysis (DDA). They explain how DDA can support the search for missing persons through statistical inference, geospatial tools, and ML and AI models. Many of the examples in the study are still at the experimental stage and have not been implemented in real-life scenarios. However, they explore the work of the Human Rights Data Analysis Group in Colombia,11 which has used ML models to combine large datasets 10 https://namus.nij.ojp.gov/ 11 https://hrdag.org/ 39 on disappearances and human rights violations. The dataset, sourced from 44 sources, contains 24 million raw records generated during the 50-year conflict in Colombia. It is playing a central role in the truth and reconciliation process (HRDAG, 2023) and demonstrates the power of AI tools to analyse and interrogate massive data sets. As part of the research, we spoke to both Mercedes Salado Puerto and Luis Fondebrider, who work with, or have worked with, EAAF. EAAF advocates for databanks, rather than more limited databases which may be constrained to one evidence type e.g., DNA. This approach has been applied, with success, to the issue of missing migrants on the USA: Mexico border (EAAF, 2023). Luis explained the kind of data they include: “From classical antemortem data, from personal belongings, from clothing, from context and all that kind of data. I'm going back to the idea to compare all the data from that information usually provided by family, hospitals and friends and the other sidethe information about the case where the body appeared, when it appeared, in what condition, with whom etc. And to see the true context together, and start analysing.” He described a hypothetical case in which bodies might be recovered after a shipwreck. If those bodies all have the same facial scarring, it could be related to a particular country, religion, or ethnic group. This could be valuable information to contribute to an identification but: “That informationthe police don't use it and they don't know how to use it because the police don't understand the concept of ethnic group. The difference is something only, in my opinion, a social anthropologist can understand, and we have the training to get that data. But you need to go to the country of origin to get that data because it's a group which is in a small village in Senegal. If you don't know the village, you don't go to the local hospital, talk with a doctor who attends to these people, who knows about the age of the people, better than people in Europe [...] It is very difficult to have information about this missing person. You will have a body, you can do any kind of technology, DNA, CT scan, but you don't have the work to compare with.” Relying solely on technical data, such as DNA profiles or CT scans of the bodies, without gathering broader, contextual information will limit identification efforts. Mercedes, elaborated about how this, more inclusive, methodology was applied in the context of the Border Project: “Forensic databanks with mixed state and non-state actors, with formal agreements and memorandum of understanding. That are working as a network nationally with transnational sharing in the collection of information in the centralization of information, comparison and information communication with the families. This is the Border Project.” For databases to work, especially at an international level, there is a lot of groundwork that needs to be done first, including standardisation of data and careful thought about data sharing mechanisms. If data on missing migrants is collected, stored, and processed in databases, data protection and security are significant concerns. Similarly, transparency about what data is being used for is crucial. Any data collected and held in Europe is covered by the General Data Protection Regulation. However, even with these stringent data protection laws in place in Europe, there are examples of migrationrelated data being unlawfully shared with other agencies, including law enforcement. For instance, EUROMED RIGHTS - A study on the use of new technologies in the identification and search for missing or deceased persons in migration in 2024, the European Data Protection Supervisor reprimanded Frontex for unlawfully transferring the data of more than 13,000 people to Europol, where it was stored and used in investigations by police authorities across Europe (Solomon, 2025). In addition, in 2022, the ICRC suffered a data breach when “servers hosting personal data belonging to more than 515,000 people worldwide were hacked in a sophisticated cyber-attack” (ICRC, 2022). This included data from the families of missing people. Florian von König (ICRC), went on to explain the delicate balance between data protection without actually over restricting what can be done with information collected: “The devil is in the detail on these things and again, more often than not, you end up discussing data protection questions which are tedious and cumbersome but important. But the more we minimise risk we may also end up restricting our ability to do anything with the information.” In terms of existing mechanisms which could be leveraged for the cross-border sharing of data relating to missing migrants in Europe, Florian von König (ICRC) also suggested Prüm Convention: “The European Prüm system is very interesting because it's a proof of concept that states are already perfectly capable of comparing different types of information. So here we're looking at the DNA, we're looking at fingerprint and other databases across Europe. And it's essentially follows what I would call querying without sharing logic.” In 2024, the Prüm legislation was updated to include information from missing persons cases. However, critics have questioned the ethics of exchanging data from unidentified human remains under a criminal identification mechanism (Machado et al. 2022). Not only does this dispersion of databases jeopardise their reliability (with false entries, duplications, and substantial information gaps), but, most importantly, CSOs, especially those based in countries of origin and transit, have expressed concerns about their limited access to these diverse databases on migrant disappearances. For instance, Ibrahima Konate, a Senegalese migration expert and founder of Missing Voices (REER) highlights that there is a significant accessibility challenge to all these databases mentioned above because the communities involved do not predominantly speak English, French, or Spanish; languages such as Wolof and Mandingo are more common. He therefore emphasises the importance of developing platforms that can communicate in these local languages. To tackle this issue, his organisation is actively developing tools to help families search for missing migrants. The organisation attempts to link communities and family collectives transnationally across Mali (via Gao and Timbuktu), Senegal (Tambacounda and Kolda) and Mauritania (Kaédi). They are also connected with diasporic communities in Europe, such as the organisation Ragazzi bayefall, based in Palermo, Italy.12 These connections facilitate the gathering of information on missing migrants, extending beyond those lost at sea to include individuals who may have been detained in Europe or elsewhere along migration routes. Increasingly, families of missing migrants are engaging in collective forms of organisation to locate their relatives. The networks through which such searches are conducted remain largely informal, yet they constitute vital infrastructures of information exchange and mutual support. Enhancing these informal mechanisms could strengthen families’ capacity to exercise their right to search and to claim recognition within broader humanitarian and state frameworks. Finally, grassroots organisations such as Missing Voices contribute to building transnational connections that transcend the Africa–Europe migration corridor, as Ibrahima Konate stated: 12 https://www.facebook.com/ragazzibayefall/ 41 “Once the platform is set up, we will also insert data from the community of missing mothers in Mexico. We will add some of their information to this platform and also send them our data. More and more people are leaving Senegal and West Africa. Even North Africans come as far as Mauritania or Senegal to take the flight and go directly to Nicaragua. So the idea is also to put information concerning sub-Saharan migrants who have left for Latin America and who have disappeared ever since.” Overall, there is a pressing need for databases to be accessible to migrant families, ensuring that a single entity does not own the information or the individual, that the data is accurate, and that privacy rights are fully respected. 12. Access Now's Digital Security Helpline Access Now’s Digital Security Helpline is an essential service that could be particularly valuable for actors working in the field.13 The helpline provides timely information and resources to individuals and organisations worldwide, helping them stay safe online and supporting the implementation of effective digital security practices. Discussions about establishing a cloud-based database or a network of locally self-built databases for missing and deceased migrants frequently arise within the sector. These conversations often revolve around security and privacy concerns, both for the tracing actors (such as human rights defenders, local authorities, migrants’ families, and researchers) and for the missing persons themselves, whose rights (including, potentially, the right to go missing) remain a significant ethical challenge. While this report does not aim to propose an ideal solution, it is essential to note that the large number of existing databases worldwide and the diverse ways in which information is encoded make it challenging to conduct efficient searches for missing migrants. New technologies offer encryption mechanisms that can protect against data interception and could help develop a more accessible and efficient system capable of linking existing local databases. However, balancing strong encryption with user-friendliness remains a significant challenge. As Giulo Coppi, Senior Humanitarian Officer with Access Now, explained, referring to specific systems: “Systems are not like blockchain, but distributed systems that are basically encrypted and pinging with each other. If you breach one, you're not getting the whole picture. [...] Small NGOs, often bootstrapped and volunteer-based, would struggle to guarantee the required level of protection and access. These risks create a form of gatekeeping, where only national-level or large international agencies can operate such systems.” During the interview, he also mentioned the ICRC’s Restoring Family Links (RFL) database, which had previously been subject to a security breach. This incident highlighted the need to consider, depending on the threat model, distributed alternatives to the traditional centralised model, ensuring that no single actor can access the full database. Yet, such a configuration for the RFL database would hardly be “compatible with the current governance setup, which is centred around the Central Tracing Agency as the core organiser and coordinator” (Interview with Giulo Coppi from Access Now). While technical solutions clearly exist, real challenges lie elsewhere: “The problem is not only technological - it’s about governance. Who will set it up? Who will conduct the technical audits? Who will manage the system and cover its operational costs? Who will provide the training? I don’t see it as a tech problem; I see it as a governance problem, as is very often the case.” (Interview with Access Now, Giulio Coppi) 13 https://www.accessnow.org/help/ EUROMED RIGHTS - A study on the use of new technologies in the identification and search for missing or deceased persons in migration 13. Social media platforms as spaces for memory, commemoration, resistance, and mourning Social media platforms – such as Facebook pages, Instagram profiles, and blogs – have also been extensively used as spaces for memory, commemoration, and the creation of grassroots memorials. Several organisations have developed their own archives documenting cases of missing migrants by gathering narratives, testimonies, and personal information directly from families and communities searching for their loved ones. One example is the Missing at the Borders database,14 created by grassroots organisations based on both shores of the Mediterranean Sea (including Milano senza Frontiere, Palermo senza Frontiere, Como senza Frontiere, Carovane Migranti, Association des Travailleurs Maghrébins de France, Alarm Phone and Watch The Med). Their mission is to join forces with the families of migrants who died, disappeared, or were subjected to enforced disappearance during their journey to Europe. The database aims to give voice to these families by documenting migration-related disappearances through testimonies, photographs, family accounts, and information on the time and place of disappearance. Relying on crowdsourced data collected from NGOs, families, and local networks across North Africa, West Africa, and the Central Mediterranean, the platform records the stories and names of those who have gone missing and, in doing so, seeks to remind the world of their individuality, hopes, and dreams. During our interview, Edda Pando from Milano Senza Frontiere and Missing at the Borders emphasised the importance of creating spaces where relatives can narrate, in the first person, the stories of their missing loved ones and their own pursuit of truth and justice. As she explained, this act of testimony has a cathartic value for families, offers a model of action for others facing similar circumstances, and – most importantly – carries a political significance: enabling those who are victims of border regimes, even indirectly, to represent themselves and assert their own political subjectivity without external mediation. Similarly, the MemMed website pursues related objectives.15 In addition to publishing reports on shipwrecks and the association’s efforts toward truth and justice, it includes a “memory” section dedicated to “the stories of people on the move who crossed the Mediterranean and lost their lives in shipwrecks, pushbacks at sea, or situations of detention and deprivation of liberty (such as hotspots and detention centres). These life stories are collected together with the families and friends of the people on the move whom the association supports in their demands for truth and justice” (author’s translation, Memoria Mediterranea 2024). The growing tendency to intertwine the commemorative dimension of deaths at sea with the collection of data and stories of the dead and missing has also transformed some border-victim databases into genuine artistic expressions. A significant example is the diverse artistic reinterpretations, created on the occasion of various commemorations, of the UNITED List of Refugee Deaths (2025),16 which has been collecting reliable data on refugee deaths related to “Fortress Europe” since 1993. The convergence of data collection, memorialization, and activism finds one of its most explicit expressions in the CommemorActions initiative led by Alarm Phone,17 both in physical protests and in its online documentation. Emerging from collaborations between families of those missing in the 14 https://missingattheborders.org/en/ 15 https://memoriamediterranea.org/en/ 16 https://unitedagainstrefugeedeaths.eu/about-the-campaign/about-the-united-list-of-deaths/ 17 https://alarmphone.org/en/2023/05/30/commemoraction/ 43 Mediterranean and activists who collect their testimonies, these actions honour those who died or disappeared while demanding justice. Combining artistic performance, political messaging, and commemoration, CommemorActions (2023), a ‘weapon of the weak’, create platforms that connect grieving families with broader publics, allowing distant relatives to participate and creating online memorials for those who are no longer present. Figure 7. Commemoraction – Alarm phone - https://alarmphone.org/en/2023/05/30/commemoraction/ Figure 8. Association Marocaine d'aide aux Migrants en Situation Vulnerable - Facebook page Last but not least, social media platforms can also function as spaces of mourning. The Association for Helping Migrants in Difficult Circumstances Oujda (Association d’Aide aux Migrants en Situation Vulnérable Oujda, AMSV),18 established in 2017 in Morocco, does extraordinary work in supporting families and ensuring that deceased migrants are buried with dignity. During our interview, Hassane Ammari, the president of AMSV, explained that, with the families’ consent, the association’s Facebook page (See Figure 8) can be used to enable relatives to attend the burial of their loved ones virtually. This practice allows dispersed family members to participate in mourning rituals despite the geographic distances. 18 https://www.facebook.com/AMSV.Oujda?locale=en_GB V. ETHICAL CONSIDERATIONS EUROMED RIGHTS - A study on the use of new technologies in the identification and search for missing or deceased persons in migration 1. The lack of a regulatory framework, more than a technical discussion Before introducing new technologies and discussing ethical considerations, there is the question of purpose and utilisation. As studies already conducted on the phenomenon of missing and deceased migrants along the migratory route that also crosses the Mediterranean demonstrate (Mediterranean Missing 2016; MemMed 2025; EuroMedRights 2025), the leading cause for the lack of search and identification of deceased and missing migrants is not technical in nature, but both political and legal. Despite the extensive legal literature that imposes on countries the obligation to identify migrants who have died at sea, encapsulated, for example, in the Mytilini Declaration (2018), states do not acknowledge the international legal obligation to identify. The lack of a regulatory framework that allows this mission to be carried out, otherwise considered humanitarian, and also justifies the expenditure to the court of auditors, is discussed by Chief Prosecutor Salvatore Vella, currently serving at the Prosecutor’s Office of Caltanissetta and previously at the Prosecutor's Office of Agrigento, which also covers the Island of Lampedusa. "Because the Public Prosecutor's Office acts when there is a crime. If there is no crime, having a dead person is not our responsibility. Even a dead person in a criminal investigation - the identity of a dead person has importance on certain occasions. That is, if you have a tragedy at sea where 200 people die, in my investigative activity, I don't have the need to understand or know who these people were - in the sense of what their names were, where they were born, why they were there. [...]It's clear that if I know it or can know it, that's better, but if I don't know it, it doesn't change anything for me in reconstructing the facts.” 45 Therefore, the primary ethical consideration in deploying new technologies to search for and identify missing persons along migratory routes is establishing a regulatory framework. Such a framework must enable judicial authorities, civil authorities, victims' families, and developers to operate within a structure where their efforts are not merely voluntary but are grounded in legislation that mandates the pursuit of truth and justice for the missing and the deceased. 2. Guiding principles around AI tools Beyond data security, as mentioned above regarding databases, there is an emerging framework of regulations and guiding principles for AI tools. These may help developers and users navigate the complex ethical and legal issues surrounding the application of AI tools in the missing migrants space and ultimately determine whether we can trust them. In 2024, the European Union implemented the AI Act (Regulation (EU) 2024/1689). It is the first-ever legal framework on AI, addressing AI risks and fostering trustworthy AI in Europe. However, as highlighted by Caterina Rodelli (Access Now), many human rights organisations have criticised the EU AI Act. As it stands, it ‘falls short in the vital area of migration, failing to prevent harm and provide protection for people on the move’ (see Joint Statement, #ProtectNotSurveil Coalition, 2024). In its current form, the legislation establishes a separate legal framework for the use of AI by law enforcement, migration control, and national security authorities. This differentiated regime creates the conditions for expanded and more harmful surveillance, discrimination, and violence against people on the move. We spoke to Óscar Ibáñez Panizo from Panacea Research Cooperative, who explained some of the principles they try to incorporate in their work, a lot of which stems from the concept of trustworthy AI under the EU AI Act, such as inclusivity, explainability and reliability: “Trustworthy AI is pushing the developer to think about safe, robust tools, explainable tools. [...] So for example, human oversight or agency or supervision, in risky contexts, it's something you have to consider. It should be self-explainable. Ideally the expert should be able to understand the why the solution is this and not this one. But not only the expert, ideally the tools you develop should also communicate with families, with other stakeholders and you have to consider also how we communicate the return.” To engender trust in AI systems, rigorous verification and validation of tools are recommended. In addition, another aspect of trust in AI systems concerns transparency and what is happening inside the system (Russell & Norvig, 2022). This transparency can be addressed through the concept of explainable AI (XAI). XAI systems can explain themselves and their outputs to humans. In fact, an AI explanation may be better than a human explanation, but it should be made clear which is which (ibid). As Óscar highlighted, it is not yet fully clear where the responsibility for the reliability of an AI tool lies, whether this is with the developers or with the practitioner who decides to use the tool. Therefore, caution is urged, especially regarding what is known about how ML and AI can perpetuate societal biases. ML algorithms give better accuracy with more diverse training data. Therefore, minorities will experience lower accuracy (Russell & Norvig, 2022). Designers of ML systems should ensure fairness by incorporating a representative range of training data. Óscar explained how expensive this can be: “Not having bias in the algorithms and the software is really costly because it means ideally we should get data from many different populations. [...]Most of the technologies are developed in the USA or in the European Union, and in the case of identification they are applied to people from the South, so it's completely different populations.” EUROMED RIGHTS - A study on the use of new technologies in the identification and search for missing or deceased persons in migration This cost, then raises the question of the role of commercial companies, who may be the only actors with the resources required for this work, as Florian von König, from the Central Tracing Agency at the ICRC outlined: “It's a fact when it comes to AI, like it or not, that the sums that are necessary to develop some of these systems and models [...]whether you like it or not, a lot of this stuff will be in the domain of commercial actors and we'll need to think long and hard whether there are acceptable ways to engage with them, because otherwise we deprive ourselves of some potentially very performant tools.” However, as Luis Fondebrider (Independent Forensic Consultant, former EAAF/ ICRC) noted, tensions often arise between commercial interests and scientific rigour. He observed that companies developing forensic technologies operate within a competitive market, where financial incentives may outpace evidentiary validation. As he put it, “companies want to earn money… but they should be honest and say the machine is ready or not ready, and needs to be tested.” Fondebrider highlighted how both private companies and universities contribute to a “jungle” of products and claims, making it difficult to discern which tools are scientifically robust and which require further research. This dynamic, he warned, also affects the DNA technology sector, where practitioners can “get lost” amid unverified or insufficiently tested innovations. Those working at the intersection of software engineering and humanitarian applications may consider themselves “humanitarian engineers”. This is a relatively understudied group of actors in the field of forensic humanitarian action, with particular relevance to new technologies. Their work warrants further exploration, particularly regarding ethics, responsibility, and accountability when their tools are used to search for and identify missing migrants. 3. The Rights of Families of the Missing and the Use of New Technologies Throughout the research, we could appreciate how some technological tools are being developed with families' needs in mind, while others are being created in total isolation. Indeed, families are much more involved in the search for missing migrants than in the identification of migrant bodies. We argue that any tools developed in this field should be conceived, designed, and implemented with the “right to search and identification” that migrant families should be able to exercise. This entails, for example, as Gabriella Citroni taught us, that in work related to disappearances (both regarding migrants and enforced disappearances), it is essential to adopt the presumption-of-life approach (2025). Following our research interviews, to follow a more holistic approach, discussions about integrating different types of databases should take place to provide families with more inclusive access to information and to exercise their "right to search." Families often carry unique forms of knowledge – such as details about personal items, social networks, migration routes, or last-known contacts – that can prove invaluable for tracing and identification. Yet their involvement has historically been limited by institutional gatekeeping, lack of transparency, and fragmented communication between humanitarian and activist actors, governments, and civil society organisations. A family-centred approach requires viewing families not as victims or passive beneficiaries but as co-producers of knowledge, partners, and rights-holders, including with respect to the right to search and the “right to identification” (Jørgensen, 2024). 47 An approach that places families at the centre therefore recognises them as primary stakeholders, as holders of essential data, but also as actors with agency who can reshape the direction of investigations – as the early cases of the Mothers of the Plaza de Mayo in Argentina teach us, and as numerous instances of forensic family-led actions continue to demonstrate (e.g., Cruz-Santiago 2020). Family participation in investigations and in the co-construction of identification models also has the advantage of producing outcomes that are more likely to be accepted by families themselves. As reported in an interview with a member of Border Forensics: “A member of a family collective told me that they do not believe in, nor trust, the identification tool. A closed coffin arrives back home—how can they be sure it is their son? A DNA test that no one has explained to them?” Therefore, involving families is not only ethically appropriate but essential for creating procedures that make identification acceptable to all parties involved. As discussed in this report, technological innovation has transformed the field of search and identification of missing migrants. Digital social media platforms, Open Source Intelligence, biometric databases, DNA matching, technologies like facial recognition, and blockchain technologies have the potential to provide more responses to families. Yet these tools also raise complex ethical questions about privacy, consent, and data ownership, particularly when dealing with highly sensitive personal and familial information. Therefore, a family-centred approach requires that new technologies be developed and deployed with families’ rights and agency in mind. VI. CONCLUSION EUROMED RIGHTS - A study on the use of new technologies in the identification and search for missing or deceased persons in migration In this report, we explore various tools and methodologies categorised as new technologies. It is important to remember that many organisations working on the issue of missing migrants and the identification of bodies often lack sufficient resources. Currently, most CSOs, communities, and collectives, including those of families of missing persons, tend to organise informally and rely heavily on the goodwill of individuals for support. In these contexts, discussions about the use of new technologies, including AI, are rarely prioritised. When new technologies were used, we mapped their potential harms and misuse, and highlighted positive practices and possible benefits. When engaging with the latest technologies, it is crucial to examine their origins, intended beneficiaries, and, in particular, how these tools can be utilised by the families of missing migrants and the civil society organisations that support them in their search for loved ones. Throughout our interviews, we learned about research projects, ideas, and tools under development related to new technologies and their potential applications for identifying and locating missing or deceased persons in migration. Additionally, we discovered a growing community of practice focused on this topic. However, the process of testing and validation can take many years when done correctly, and there currently appear to be very few operational tools available for daily use by forensic practitioners, humanitarian workers, civil society organisations, or families. It seems inevitable that new technologies and AI will make their way into this field and contribute in meaningful ways. Therefore, this moment presents an opportunity to analyse the usefulness of these tools and the costs associated with their use. We aim to address these concerns by providing a critical evaluation of technology's role in this specific context. We emphasise the crucial ethical considerations surrounding new technologies in the context of missing migrants. Regulations and guiding principles are essential to mitigate potential harm, and the rights of the families of the missing must always be prioritised. Ultimately, we conclude that new technologies are neither inherently good nor inherently bad; rather, their value depends on how they are used, governed, and the ethical frameworks surrounding them. BIBLIOGRAPHY 49 Alarcón,Á., Giandana Gigena, F. and G. Coppi (2024). “Smart borders” and the making of a humanitarian crisis. Access Now. 13 March 2024. https://www.accessnow.org/smart-borders-and-the-making-of-a-humanitarian-crisis/ Amankwaa, A.O., Kupčuláková, B. and Randall, K. (2025) ‘Humanitarian Forensics: Stakeholder perspectives on the use of international kinship matching DNA databases’, Journal of Forensic and Legal Medicine, 116, p. 103018. Available at: https://doi.org/10.1016/j.jflm.2025.103018 Anderson, B.E. (2008) ‘Identifying the Dead: Methods Utilized by the Pima County (Arizona) Office of the Medical Examiner for Undocumented Border Crossers: 2001–2006’, Journal of Forensic Sciences, 53(1), pp. 8–15. Available at: https://doi.org/10.1111/j.1556-4029.2007.00609.x Assembly debate on 3 October 2025 (36th sitting) (Doc. 16240, report of the Committee on Migration, Refugees and Displaced Persons, rapporteur: Mr Petri Honkonen). Text adopted by the Assembly on 3 October 2025 (36th sitting). (https://pace.coe.int/pdf/123f34b17161067a041a87ea8a7d9d272492e45d84ea406798f96c69897516ef/res.%20 2628.pdf) Barash, M. et al. (2024) ‘Machine learning applications in forensic DNA profiling: A critical review’, Forensic Science International: Genetics, 69, p. 102994. Available at: https://doi.org/10.1016/j.fsigen.2023.102994 Baraybar, J. P., Caridi I. and Stockwell J.(2020). “A Forensic Perspective on the New Disappeared.” In Forensic Science and Humanitarian Action, edited by R. C. Parra, S. C. Zapico, and D. H. Ubelaker. Available at: https://doi:10.1002/9781119482062.ch7 Bennett, C. (2014) ‘Who knows who we are? Questioning DNA analysis in disaster victim identification’, New Genetics and Society, 33(3), pp. 239–256. Available at: https://doi.org/10.1080/14636778.2014.946003 Bhuiyan, J. (2022) ‘Ukraine uses facial recognition software to identify Russian soldiers killed in combat’, The Guardian, 24 March. Available at: https://www.theguardian.com/technology/2022/mar/24/ukraine-facial-recognitionidentify-russian-soldiers Blau, S. et al. (2023) ‘Re-examining so-called “secondary identifiers” in Disaster Victim Identification (DVI): Why and how are they used?’, Forensic Science International, 345, p. 111615. Available at: https://doi.org/10.1016/j.forsciint.2023.111615 Bowman, Z. et al. (2022) ‘Rapid DNA from a disaster victim identification perspective: Is it a game changer?’, Forensic Science International: Genetics, 58, p. 102684. Available at: https://doi.org/10.1016/j.fsigen.2022.102684 Boss, P. and Yeats J.R.(2014). “Ambiguous Loss: A Complicated Type of Grief When Loved Ones Disappear.” Bereavement Care 33 (2): 63–69. Available at: http://doi.org/10.1080/02682621.2014.933573 BVMN (2023) Border Violence Monitoring Network, ‘EU Member States' use of new technologies in enforced disappearances: Input for the thematic study by the UN Working Group on Enforced or Involuntary Disappearances on “new technologies and enforced disappearances". Available at: https://borderviolence.eu/uploads/document/file/353/ Input-for-the-thematic-study-on-new-technologies-and-enforced-disappearances_version-2.pdf Caplova, Z. et al. (2017) ‘The Reliability of Facial Recognition of Deceased Persons on Photographs’, Journal of Forensic Sciences, 62(5), pp. 1286-1291. Available at: https://doi.org/10.1111/1556-4029.13396 Cattaneo, C. et al. (2022) ‘The rights of migrants to the identification of their dead: an attempt at an identification strategy from Italy’, International Journal of Legal Medicine, 137, pp. 145–156. Available at: https://doi.org/10.1007/s00414-022-02778-1 Chong M, Sheehan S, Battaglia J, Wescott DJ, Wallin J (2023). Comparative study of Rapid DNA versus conventional methods on compromised bones. Forensic Sci Int Genet;63:102825.Available at: http://doi.org/10.1016/j.fsigen.2022.102825 Citizen Evidence Lab. (2025). Amnesty International Evidence Lab. Harnessing digital methods for human rights