Artificial Turn Migrations and Asylum at the Encounter with Safe Countries of Origin in The Ontologies of Borders and the Epistemologies of Control

This article explores the Artificial Turn in European migration governance, where artificial intelligence, digital infrastructures, and dual-use technologies redefine the legal and epistemic boundaries of asylum and border control. Drawing on the aftermath of the 2015 migration crisis and the 2024 New Pact on Migration and Asylum, it examines how algorithmic systems and data extraction practices - particularly those targeting migrants’ biometric and digital data -reshape notions of “safe countries of origin” and transform the relationship between protection and surveillance. Through a pilot comparative study involving GPT-5 and Chat DeepSeek-R1, this paper illustrates how AI systems reproduce inconsistencies and normative ambiguities when classifying countries as “safe” thereby challenging human rights standards and the principle of non-refoulement. The analysis reveals how dual-use technologies blur the boundary between humanitarianism and security, accelerating the automation of credibility and identity assessments while eroding procedural safeguards. The paper calls for a human-rights-based approach to AI deployment at borders—grounded in transparency, judicial oversight, and interpretative accountability—to ensure that the governance of migration in the digital age and datafication process remains faithful to the rule of law and human dignity.