THE LEGAL LEGITIMACY OF ARTIFICIAL INTELLIGENCE AND AUTOMATED DEPORTATION SYSTEMS

The integration of artificial intelligence into automated deportation systems has transformed border governance, presenting both operational efficiencies and significant human rights challenges. This study examines the technical infrastructure, legal implications, and ethical concerns surrounding AI-driven deportation mechanisms through comparative analysis of systems deployed in the EU, U.S., and Greece. Technically, these systems combine biometric identification (facial recognition, fingerprint analysis), machine learning-based risk assessment, and automated decision-making tools. While promising faster processing and consistent application of rules, they introduce critical vulnerabilities in legal compliance and fundamental rights protection. Our research reveals three core legitimacy deficits: (1) violations of non-refoulement principles through automated decisions lacking individualized assessment; (2) erosion of due process rights due to opaque algorithmic processes; and (3) discriminatory outcomes stemming from biased training data and proxy variables. The ethical analysis demonstrates how these systems institutionalize structural inequalities. First, the "black box" nature of algorithms prevents meaningful scrutiny of deportation decisions. Second, expansive biometric data collection occurs without proper consent protocols. Third, cross-border data sharing enables mission creep beyond original purposes. Case studies of U.S. deportation prioritization systems and EU border control projects illustrate these systemic risks. Key findings indicate that current implementations frequently conflict with international human rights standards (Nuredin, 2022b), particularly under the ECHR and Refugee Convention. Automated systems often fail to meet proportionality requirements, lack necessary safeguards against errors, and undermine the right to effective remedy. The displacement of human judgment to algorithms in life-impacting decisions raises fundamental questions about justice in migration governance. The study proposes a regulatory framework centered on four principles: (1) mandatory human rights impact assessments for all migration AI systems; (2) legally enforceable explainability requirements; (3) independent oversight mechanisms with technical auditing capacity; and (4) international harmonization of standards through UNHCR guidance. These findings suggest that without urgent intervention, the current trajectory of automated deportation systems risks normalizing rights violations under the guise of technological efficiency. While AI can potentially improve certain aspects of migration management, its use in deportation contexts requires stringent safeguards to prevent the automation of injustice and preserve constitutional protections in increasingly digital border regimes.