Working with AI at the gates?: Insights from the Perception of Refugee Officers

• In Event: Insights on Artificial Intelligence in Government

Abstract

The integration of artificial intelligence (AI) into public sector operations presents significant challenges and opportunities, particularly for frontline workers who engage directly with citizens. Particularly, forging collaboration between human and AI is one of big questions in the transformative and resilient policy era. This study investigates the perceptions and management strategies of 106 frontline workers in South Korea who handle asylum seekers, focusing on their perceptions, understanding and cautions of AI technology. Utilizing Bullock’s (2019) typology, which contrasts the strengths and weaknesses of human versus AI activities under varying levels of uncertainty and complexity, this study explores whether public officers' perceptions align with these theoretical insights.

Through comprehensive surveys, we assessed the front-line workers' views on the complexity and uncertainty of each task and their opinions on AI adoption for these tasks. The research categorizes the task about decision-making process on asylum seekers into five key steps: (1) receipt and classification of application documents, (2) assessment of the application documents, (3) information collection regarding the source country, (4) interviews with refugee applicants, and (5) decision-making. Results indicate that interviewing refugee applicants is perceived as the most complex and uncertain task, followed by decision-making. Conversely, the receipt and classification of application documents were considered the least complex and uncertain. Officers showed the greatest support for AI adoption in information collection regarding source countries and, secondarily, in document receipt and classification. They viewed interviews and decision-making as tasks least suitable for AI replacement. While officers acknowledged AI's potential to enhance efficiency and reduce errors in documentation processes, they expressed concerns about AI's role in decision-making, emphasizing that AI should supplement rather than replace human judgment. Transparency and open access to the algorithms and data underpinning AI decisions were deemed essential.

This study contributes to the understanding of AI's role in public management by providing empirical insights into frontline workers' perspectives, highlighting both the potential benefits and ethical considerations of AI integration in public service settings.

Author

• Do Han Kim, Kyung Hee University, South Korea
  Presenting Author