IRCC immigration and visa application AI screening

Released: 2018

Can you improve this page?
Share your insights with us

Immigration, Refugees and Citizenship Canada (IRCC) has been running an 'advanced data analytics' and artificial intelligence (AI) pilot since 2018 to process temporary resident visa (TRVs) applications submitted from China and India. The system was extended to TRV applications submitted from all countries outside Canada in January 2022.

Visa applications are assessed on the basis of eligibility and admissibility. For straightforward applications, eligibility is approved solely by the model, while eligibility for more complex applications is decided upon by an immigration officer. All applications are reviewed by an immigration officer for admissibility.

Reaction

A September 2018 report (pdf) by University of Toronto's Citizen Lab concluded the IRCC's pilot threatens to violate domestic and international human rights law in the form of bias, discrimination, privacy breaches, due process, and procedural fairness.

The report cautions that experimenting with these technologies in the immigration and refugee system amounts to a 'high-risk laboratory,' as many of these applications come from some of the world’s most vulnerable people, including those fleeing persecution and war zones. 

Citizen Lab called on the national authorities to freeze the development of AI-based systems until a government standard and oversight bodies are established. 

In January 2022, the IRCC published an algorithmic impact assessment (AIA) of its temporary resident visa application assessment system, concluding that the impact level of the system is moderate. 

However, the assessment has been described as 'noticeably short and trite in detail', noticeably regarding how its scoring system works, the identity and degree of involvement of external stakeholders, and transparency.

Transparency

The IRCC's TRV application assessment system has been shrouded in secrecy since its inception. 

During the Citizen Lab report’s research phase, 27 distinct official information requests were submitted to the Government of Canada. Every one of them remained unanswered.

And while the publication of the IRCC's algorithmic impact assessment provided further information about the system, many questions were limited to 'yes' or 'no' answers.

Operator: Immigration, Refugees and Citizenship Canada (IRCC)
Developer: Immigration, Refugees and Citizenship Canada (IRCC)
Country: Canada
Sector: Govt - immigration
Purpose: Process temporary resident visa applications
Technology: Data analytics; Machine learning
Issue: Privacy; Bias/discrimination - gender, race; Fairness
Transparency: Governance; Black box; Marketing

System

Research, advocacy

News, commentary, analysis