Skip to content
This is NOT an official site of the Government of Canada. Click here for the official AI registry.

Helpful Places

Government of Canada – DTPR AI Register

AI-Assisted Case Prioritization for Old Age Security Compliance

Eligibility & Public Benefits · Enforcement

What it collects that can identify you

Sensitive personal information
Identifiable data
  • OAS recipient personal information from the ESDC OAS pensions system and SIN registry, including identity details, benefit status (active, ineligible, etc.), and recipient state (alive or deceased). Data is Protected B and linked to identified individuals.
About a measurement
Identifiable data
  • Traveller entry and exit records from the CBSA traveller information system, used to calculate the length of an OAS recipient's absence from Canada. These records are keyed to identified individuals via SIN or identity matching.
Run by
Employment and Social Development Canada (ESDC)
Where
No fixed location
Kept
Retained Not specified in the AIA; governed by applicable Personal Information Banks (CBSA PPU 1202, ESDC PPU 116, ESDC PPU 336, ESDC PPU 390)
Shared with
Not stated by the Helpful Places.
Your copy
You cannot see the data it holds about you. What you can do

What it is for

Eligibility & Public BenefitsEnforcement

This system uses AI clustering techniques to help government investigators identify and prioritize Old Age Security (OAS) recipients who may have been absent from Canada for more than six months without reporting it — a condition that affects eligibility. The system groups cases by priority but does not make any eligibility decisions itself; all decisions are made by human investigators. Individuals flagged by the system are given an opportunity to provide documentation before any determination is made.

What it collects and what happens to it

Data taken in

Sensitive personal information
Identifiable data
  • OAS recipient personal information from the ESDC OAS pensions system and SIN registry, including identity details, benefit status (active, ineligible, etc.), and recipient state (alive or deceased). Data is Protected B and linked to identified individuals.
About a measurement
Identifiable data
  • Traveller entry and exit records from the CBSA traveller information system, used to calculate the length of an OAS recipient's absence from Canada. These records are keyed to identified individuals via SIN or identity matching.

Processing

Clustering & Segmentation
  • A self-organizing map (SOM) and clustering model group OAS recipient cases based on (1) length of absence from Canada, (2) recipient state (alive or deceased), and (3) OAS benefit status. Cases are labelled and summarized by cluster to support investigator triage.

What it does

Deciding (Analytical AI)
Human decides
  • The system scores and ranks OAS cases by priority using clustering criteria (length of absence, recipient status, benefit status). It produces a prioritized list and grouped summary for investigators, who then decide whether to open a formal inspection. No eligibility decision is made by the system.

Outputs

A recommendation or prediction
Anonymized data
  • The system outputs filtered observations labelled and summarized by group, providing investigators with a prioritized list of cases requiring immediate assessment. This is advisory only — investigators conduct a manual pre-assessment before any inspection begins and retain full decision authority.

Run by

Employment and Social Development Canada (ESDC)
  • Service Canada, part of Employment and Social Development Canada, deploys and operates this system through its Integrity Service Branch to detect misuse of the Old Age Security program. The project is conducted with oversight from the Artificial Intelligence Centre of Excellence in accordance with Treasury Board of Canada Secretariat guidelines.

OAS Extended Absence Leave Prioritization — Register Entry

Built by

Canada Border Services Agency (CBSA)
  • The Canada Border Services Agency supplies traveller entry and exit information to Service Canada, which is a key input to this system. CBSA data is used to determine the length of time OAS recipients have been outside Canada.

AIA — Section 3.1 Q53: Input data described

Kept for

Retained Not specified in the AIA; governed by applicable Personal Information Banks (CBSA PPU 1202, ESDC PPU 116, ESDC PPU 336, ESDC PPU 390)
  • The AIA does not specify a retention duration for system outputs. Retention is governed by the relevant Personal Information Banks registered with the Office of the Privacy Commissioner of Canada. A Privacy Impact Assessment was completed in July 2020.
  • Duration: Not specified in the AIA; governed by applicable Personal Information Banks (CBSA PPU 1202, ESDC PPU 116, ESDC PPU 336, ESDC PPU 390)

Shared with

Not available to me
  • Outputs of the AI prioritization system (grouped case labels and summaries) are accessible only to Integrity Operations agents within Service Canada. The system operates within a closed network with no internet connections, and access is controlled via a grant-monitor-revoke permission process.

Stored

Stored locally
  • The system operates within a closed system with no internet or intranet connections. Data is controlled by the federal government of Canada and stored within federal infrastructure in Canada.
  • Duration: Not specified; governed by applicable Personal Information Banks
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 UseThe Algorithmic Impact Assessment is publicly available on the Government of Canada Open Data portal, providing notice that an AI-assisted prioritization system is used in OAS compliance investigations. The AIA describes the system's role, input data, and the fact that all eligibility decisions are made by human investigators.
  • Right to a Human ReviewAll inspection and eligibility decisions are made by human Integrity Operations agents. A manual pre-assessment is conducted before any inspection begins. Individuals identified by the system are given an opportunity to present documentation and explanation before investigators produce any decisions.
  • Right to ContestIndividuals who receive an administrative determination of ineligibility may request a reconsideration of the decision under section 27.1 of the Old Age Security Act.
  • Right to Algorithmic TransparencyThe Algorithmic Impact Assessment provides a plain-language description of the system's role, input data sources, criteria used to evaluate cases, and system output. It is publicly available on the Government of Canada Open Data portal. Under the Treasury Board Directive on Automated Decision-Making, a meaningful explanation of common decision results must be published and discoverable via the departmental website.

Risks and safeguards

  • Financial & business harmIncorrect case prioritization could lead to wrongful suspension of OAS or GIS benefit payments for individuals who are actually eligible, causing financial harm.Safeguard: The system does not render decisions — all cases are pre-assessed and fully investigated by human Integrity Operations agents before any payment action is taken. Individuals have recourse to request reconsideration under s. 27.1 of the OAS Act. An audit trail records all system recommendations and decision points.
  • Civil liberties harmThe clustering model could disproportionately flag or group vulnerable or at-risk OAS recipients (e.g. those with complex travel patterns due to disability or cultural practices), raising concerns about discriminatory impact and due process. The AIA notes that 'if the system were to malfunction, there is a risk it could inadvertently overrepresent at-risk recipients.'Safeguard: An ethical review and fairness assessment is planned before deployment. The system does not make binding decisions — human investigators assess each case individually and individuals may present documentation before any determination. Legal services reviewed the project under the Directive on Automated Decision-Making. A Privacy Impact Assessment was completed (Jul 2020). No Gender-Based Analysis Plus of the data was conducted, which is a noted gap.