AI Translation Errors in Immigration Courts Affected Asylum Decisions

In early 2025, the Executive Office for Immigration Review (EOIR) began utilizing AI-powered translation systems to address interpreter shortages in immigration courts across the United States. These systems were deployed to provide real-time translation services for asylum hearings involving non-English speaking applicants, particularly those speaking less common languages where certified interpreters were scarce. Research conducted by immigration advocacy organizations documented numerous translation errors that occurred during asylum proceedings. The AI systems struggled with nuanced legal terminology, cultural context, and dialectal variations that are critical for accurate communication in asylum cases. Specific errors included mistranslating descriptions of persecution, incorrectly rendering legal concepts, and failing to convey emotional testimony that is often crucial for establishing credibility in asylum claims. The impact on asylum seekers was significant, as these translation errors potentially influenced judicial decisions on life-or-death asylum cases. Immigration judges, relying on AI translations without human verification, may have made determinations based on incomplete or inaccurate information about applicants' circumstances and persecution claims. This raised serious due process concerns, as accurate translation is fundamental to fair legal proceedings. Legal advocacy groups filed challenges in federal court, arguing that the use of unverified AI translation systems violated asylum seekers' constitutional rights to due process and effective assistance. The litigation highlighted the inadequacy of AI systems for complex legal proceedings requiring cultural sensitivity and precise legal interpretation. The cases drew attention to the broader implications of deploying AI tools in judicial settings without adequate human oversight and quality control measures.