AI-Assisted Eligibility Triage for Spousal Immigration Applications

Eligibility & Public Benefits · Risk Assessment & Triage

What it collects that can identify you

Sensitive personal information

Identifiable data

Personal information submitted by sponsors and principal applicants as part of Spouse or Common-law Partner in Canada permanent residence applications. Data is classified at Protected B/Protected C and collected by IRCC.

AIA — Section 3.1, Questions 26–36

Also collects operational data, which is anonymized data.

Run by: Immigration, Refugees and Citizenship Canada (IRCC)
Where: No fixed location
Kept: Retained not specified in AIA
Shared with: Accountable organization
Your copy: You cannot see the data it holds about you. What you can do

What it is for

Eligibility & Public Benefits

Risk Assessment & Triage

This AI system analyzes applications for Spouse or Common-law Partner in Canada permanent residence to automatically approve eligibility for some sponsors and principal applicants, and to triage others for officer review. The system never refuses applications — all refusals and final decisions remain with human immigration officers. It processes personal information submitted by applicants and is used by Immigration, Refugees and Citizenship Canada (IRCC) employees.

What it collects and what happens to it

Data taken in

Sensitive personal information

Identifiable data

Personal information submitted by sponsors and principal applicants as part of Spouse or Common-law Partner in Canada permanent residence applications. Data is classified at Protected B/Protected C and collected by IRCC.

AIA — Section 3.1, Questions 26–36

Operational data

Anonymized data

Structured application data from Spouse or Common-law Partner in Canada submissions, used to identify patterns and eligibility signals. Data elements have a clear link to legislative and regulatory requirements.

AIA — Section 3.1, Question 19

Processing

Classification & Prediction

The system uses classification and prediction on structured application data to determine eligibility. Model rules were reviewed by experienced officers, legal, policy, data science, and privacy experts, as well as senior decision-makers.

AIA — Section 3.1, Question 19

Clustering & Segmentation

The system analyzes large data sets to identify anomalies and cluster patterns in application data, supporting workflow triage and process optimization.

Canada AI Register — Capabilities field

What it does

Deciding (Analytical AI)

Human executes

The model predicts and classifies applicant eligibility from structured application data, automating positive determinations. For automated positive approvals, an officer still carries out the final admissibility screening and decision.

AIA — Section 3.1, Question 13

Outputs

A decision about you

Identifiable data

Automated positive eligibility determinations for some sponsors and principal applicants. The model determines only that an applicant is eligible; the application is then sent to an officer for admissibility screening. The model never refuses applications.

AIA — Section 3.1, Question 13

A recommendation or prediction

Identifiable data

For applications where the model cannot approve eligibility, applications are triaged and routed to officers for full individualized assessment. Officers are not informed of the model's triage logic or analysis results.

AIA — Section 3.1, Question 19

Run by

Immigration, Refugees and Citizenship Canada (IRCC)

IRCC is the federal department responsible for immigration, refugee protection, and citizenship in Canada. It deploys this AI system to streamline eligibility assessments for Spouse or Common-law Partner in Canada permanent residence applications.

Canada AI Register — Department field

Built by

Not stated by the Helpful Places.

Kept for

Retained not specified in AIA

The AIA indicates audit trail records are maintained and logs of all changes to the model are kept, but the specific data retention duration is not stated in the available documentation.
Duration: not specified in AIA

AIA — Section 3.2, Questions 17–23

Shared with

Available to the accountable organization

Output data and audit trail records are available to IRCC employees (GC employees are the primary users). Access permissions are granted, monitored, and revoked through a documented process.

Canada AI Register — Primary users field

Not available to me

Applicants do not have access to the model's triage logic, internal analysis, or the specific rules used for automated eligibility determinations. Officers are also not informed of the model's triage analysis.

AIA — Section 3.1, Question 19

Stored

Not stated by the Helpful Places.

How to read the colours

Can it identify you?

Anonymized data: Data about people with the link to who is broken. Stripped of identifiers, blurred, aggregated, or noised so this system can’t reasonably tie a record back to an individual.
Pseudonymous data: Each person’s data is tied to a token (hash, ID, template) that lets this system recognise the same person across events, but the token itself doesn’t reveal a name. Reidentification is possible with extra information.
Identifiable data: The data either contains a direct identifier (name, address, account name, recognisable face or voice, plate number) or carries a token this system uses to look up legal identity during processing.

Who completes the loop?

Human decides: This mode suggests; a person decides what to do next. The AI is always advisory — a human is in the loop on every decision. Example: a triage tool ranks cases for a clinician who chooses which to see first.
Human executes: This mode decides; a person carries out the result. Example: an optimizer plans the day’s trash-collection routes, and drivers run them.
Autonomous: This mode decides and acts on its own. No person reviews each decision or carries out the resulting action.

Definitions from the DTPR standard. Amber is about your data, violet about who decides. The fuller the shape and the deeper the colour, the more identifying the data or the less a person is involved.

What you can do

Ask about this system

Questions go to the Helpful Places, not the vendor.

Your rights

Right to Be Informed of AI UseIRCC is required to post a plain language notice through all service delivery channels (internet, in person, mail, or telephone) informing applicants that an automated decision system is in use. AI use has been disclosed to users.
Right to ContestA recourse process is established for clients who wish to challenge decisions made by or with the assistance of this system. Clients can request reconsideration. All applications where the model did not approve eligibility receive a full individualized assessment by an officer.
Right to an Explanation of a DecisionA meaningful explanation must be provided with any decision resulting in denial of a benefit or service. The system is able to produce reasons for its decisions when required, and the audit trail can help generate decision notifications and statements of reasons.
Right to a Human ReviewOfficers continue to make the final decision on every application, including those where the model has approved eligibility (officers screen for admissibility). The system enables human override of automated decisions, and overrides are logged. Applications the model cannot approve receive full human assessment.

Risks and safeguards

Civil liberties harmThe system makes or influences determinations in the immigration domain — an area under intense public scrutiny with potentially irreversible long-term impacts on individuals' rights and family unity.Safeguard: The model only automates positive eligibility determinations and never refuses applications; all refusals remain with human officers. Model rules are linked to legislation and reviewed by legal, policy, and privacy experts. Officers make the final decision on every application. An audit trail records all system decisions and supports human oversight. A recourse process is established for clients who wish to challenge decisions.
Reputational harmAutomated eligibility determinations could stigmatize applicants or create unfair categorizations through model bias.Safeguard: Gender-Based Analysis Plus (GBA+) was conducted on the data. Documented processes exist to test datasets against biases. Regular monitoring and quality assurance measures are in place to identify bias or discrimination early. The system includes human override capability and a feedback capture mechanism.