Human-Centered AI Applications for Canada's Immigration Settlement Sector

AI in immigration has focused on state control rather than immigrant support: Most existing AI applications in immigration are used for selection and screening purposes that primarily benefit authorities, while their reliability remains poorly studied despite their significant impact on immigrants' lives.

Settlement phase represents an underexplored opportunity for beneficial AI: The paper identifies Canada's immigration settlement phase—where newcomers need information and integration support—as a prime area for AI applications that could directly empower immigrants rather than restrict them, yet this sector has received minimal research attention.

Information access is critical but service providers are overwhelmed: The settlement sector faces a key challenge where access to accurate information is crucial for immigrant integration, but service providers lack sufficient resources to meet demand, creating an opportunity for AI tools to fill this gap responsibly.

Human-centered design is essential for trustworthy immigration AI: The authors emphasize that responsible AI solutions in this context must be explicitly designed to be empowering, inclusive, and safe—requiring collaboration between AI researchers, service providers, and government organizations rather than top-down implementation.

Multidisciplinary collaboration is necessary, not optional: The paper calls for AI researchers to move beyond technical development and actively partner with settlement service providers and other stakeholders to ensure AI tools are tailored to real immigrant needs and contexts.