Human Decision-Making and Artificial Intelligence in Canada’s Refugee System

This paper establishes how the Canadian government is using AI in decision-making related to Canada’s immigration system to argue that decisions relating to refugees and asylum seekers are highly complex and involve great uncertainty, where discretion is important and hence, the use of AI should be avoided. To support this argument, the author develops and employs a theoretical framework for the application of artificial intelligence, highlighting the reasons, methods for and contexts in which artificial intelligence may be introduced in Canada’s refugee and asylum seeker streams. By mapping the journey of claiming refugee or asylum seeker status, this paper shows how even as intentions change, the current administration’s decision to not introduce AI in the refugee and asylum seeker streams should remain in place, as decisions made in these immigration streams are highly complex and involve great uncertainty, where discretion is important.