REFUGEE STATUS DETERMINATION THROUGH ALGORITHMIC DECISIONMAKING: JUSTICE OR DISCRIMINATION?

The integration of algorithmic decision-making systems (ADMS) into refugee status determination marks a critical evolution in global migration governance. These systems, powered by artificial intelligence and machine learning, are increasingly used to assess asylum claims, evaluate applicant credibility, and generate risk scores based on historical data. Advocates argue that such technologies can enhance efficiency, consistency, and objectivity in overburdened asylum systems. However, serious concerns persist regarding their potential to entrench systemic biases, limit transparency, and erode procedural justice. Examples from the United States, European Union, and Australia illustrate both the promise and perils of this approach. Systems like the U.S. Risk Classification Assessment and the EU’s iBorderCtrl pilot have faced criticism for reinforcing discriminatory patterns embedded in training data. Algorithms may disfavor specific nationalities or overlook credible claims that deviate from learned templates, thereby reproducing inequality under the guise of neutrality. A major ethical issue is algorithmic bias, where historically prejudiced data leads to discriminatory outcomes. The “black box” nature of many ADMS compounds this problem, as asylum seekers are often unable to understand or challenge the basis of adverse decisions. Additionally, these systems can marginalize individuals lacking digital access or literacy, deepening socioeconomic disparities within asylum procedures. Despite potential benefits such as faster processing and standardization, the equitable use of ADMS depends on robust safeguards. Many experts advocate for hybrid models that combine automated tools with meaningful human oversight. To ensure justice, comprehensive policy frameworks are essential—requiring human rights impact assessments (Nuredin, 2022b), enforceable transparency obligations, and globally harmonized ethical standards, particularly through bodies like UNHCR and OECD. Ultimately, the success of ADMS in refugee determination must be evaluated not by efficiency alone, but by their ability to uphold fundamental human rights. The future of ethical AI in migration relies on placing justice, accountability, and fairness at the center of digital innovation.