By Costica Dumbrava

 Artificial intelligence (AI) technologies are increasingly used in the areas of migration and asylum. These technologies promise to increase efficiency, reduce uncertainty and improve decision-making. In the asylum context, AI applications are meant to reduce pressure on national asylum systems and contribute to fairer and more consistent asylum decisions. Despite potential benefits, these technologies come with significant risks. Inaccurate or biased AI applications may jeopardise the right to asylum, increase or reinforce discrimination, and diminish procedural safeguards built into the asylum process. These risks depend on the type and complexity of applications used, their role in asylum procedures, and the way in which decision-makers interact with technologies. Although the deployment of AI applications in EU asylum systems is still in the early stages, the debate about the legal, social and ethical challenges posed by these new technologies is well under way. For instance, several Member States have started using AI technologies for dialect recognition to verify information, or obtain further information, on asylum applicants' country or region of origin. There are also examples of Member States using AI applications for name transliteration, automatic transcription of speeches and case matching. The introduction of AI technologies in EU asylum systems raises several major risks, relating to: (i) inaccuracies, bias and discrimination; (ii) undermining fairness and due process; and (iii) unlawful interference with privacy and data protection.

Brussels: EPRS | European Parliamentary Research Service , 2025. 12p.

By Agustina Laurito, Ashley N. Muchow 

Deportations of immigrants from the United States have grown substantially over the past two decades. While existing research has examined how changes to US laws and policies have contributed to this increase, less attention has been given to how conditions in immigrants’ countries of origin shape deportation patterns. This article investigates how an important external shock—home-country natural disasters—influences immigrant removals from the US. We combine annual data on removals by country with information on natural disasters to estimate difference-in-differences models that exploit exogenous variation in the timing and magnitude of natural disasters across countries. Our results show that immigrant removals increased, on average, by 29 percent after salient natural disasters. When we explore mechanisms, we find little evidence that home-country natural disasters increase irregular migration, but we do find that noncitizen and likely undocumented immigrants increase their labor force participation and employment following these shocks. This finding suggests that natural disasters in immigrants’ countries of origin may influence the economic behavior of immigrants, putting them at greater risk of detection by immigration enforcement authorities.