Migration Control: An AI chatbot from Frontex is to advise on self-deportation

The European Border and Coast Guard Agency (Frontex) is developing an application featuring an AI chatbot designed to advise refugees on "voluntary" return to their countries of origin. Frontex believes this application does not pose a high risk under the AI Act, despite its potential advisory role in migration processes. This initiative is part of Frontex's broader reintegration programs, which also encompass financial incentives and the facilitation of returns. The chatbot aims to inform individuals about financial support for voluntary departures and provide information on where to seek further assistance. This approach is an extension of Frontex's strategy to encourage voluntary returns, a policy that has reportedly seen success in a significant percentage of cases where "return specialists" were involved. Critics argue that for individuals with limited alternatives, such "voluntary" returns can be more accurately described as self-deportation. The classification of the AI chatbot's risk level under the EU's AI Act is a critical aspect, as AI systems used in migration contexts are often designated as high-risk. Frontex's Fundamental Rights Office (FRO) assessed the technology and concluded it does not meet the criteria for high-risk, thus obviating the need for further fundamental rights impact assessments. However, subsequent recommendations from FRO suggest theoretically substantiating the link between increased information and higher rates of voluntary returns.