Digitalization of Asylum Procedures in the EU

December 6, 2021
By Parviz Khyber

The European Commission wants to make the asylum procedures more efficient and prevent the risk of future crises and considers using new technologies. The question is, will digitalization make the asylum process more efficient or rather challenging?

The European Commission last year set up the EC Expert Group on the Views of Migrants composed of persons with a migrant background to participate in the development and implementation of migration, asylum and integration policies. When the expert group was called for the fourth meeting, the digitization of asylum procedures and the use of artificial intelligence for migration forecasting was discussed.

The Commission focused on the developments in the field of digitalization of asylum procedures for migrants to make it more efficient. The speakers indicated that digitalization will improve lodging of applications, provision of information and that it will provide a possibility to check the status of applications. It will also help to analyze languages for origin determination. Most importantly, digitalization will allow the authorities to conduct remote interviews to make the asylum process even swifter.

Sayed Parviz sitting in a sofa, by the computer. — Sayed Parviz is part of the expert group providing advice and expertise on policies in the field of migration, asylum and integration of migrants to the European Commission.

However, digitalization also raised many questions about whether it will make the asylum process more efficient or rather challenging. For instance, all migrants do not have access to the internet or a location to give interviews or they might not feel comfortable talking to a screen. It is also hard to take fingerprints or present evidence virtually. In addition, the speakers pointed out that the use of digital technologies, such as artificial intelligence for migration forecasting will reduce the risk crisis and be unprepared.

Most of the expert group, including me, perceived digitalization of asylum positively only if it truly leads to an effective and efficient asylum procedure. For instance, developing platforms which help integration and language learning, or provide legal and psychological support would be beneficial.

AI fails to detect anxiety

The experts also showed strong objections concerning the use of technologies to analyze languages for origin determination. Such technologies work on machine learning and AI, which should not be used for decision-making. Most of the refugees have either lived in migration for generations or lived in several places where they lost their original language and dialect and adopted a new one. For example, I was born in Afghanistan but since I have lived in many places, I now speak multiple languages, but cannot speak my original dialect.

This is a fact for countless other refugees and a machine cannot detect this. Additionally, AI disregards cultural differences and fails to detect anxiety, stress and trauma, which leads to an inaccurate decision.

The designing process must include migrants

Although the European Asylum Support Officer stated that the language assessment for detecting origin “only provides indication”, we expressed our disapproval and asked for its elimination because it can still affect the decision. The experts suggested that language assessment should only happen by expert humans instead of machines. In case the EU wishes to develop new technologies for language analysis or origin determination, then its designing process must include migrants.

I am Sayed Parviz Khyber. Currently I am doing my BA in Economics, Politics and Social Thought at Bard College Berlin. Since October 2020, I am a member of the European Commission Expert Group on the Views of Migrants. The group, consisting of 20 members, is appointed for a period of two years and are expected to meet two or three times per year. Each meeting lasts a full day and is held in Brussels or online, depending on the ongoing COVID-19-related restrictions.