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 Security Screening Triage for Border Officers

Border & Immigration · Risk Assessment & Triage

What it collects that can identify you

Sensitive personal information
Identifiable data
  • Referral data containing narrative text about individuals under security screening review. This data includes personal information about individuals subject to border security assessment, sourced from STS data (training) and the future Epsilon source system.
Run by
Canada Border Services Agency (CBSA)
Where
No fixed location
Kept
Not stated by the Helpful Places.
Shared with
Accountable organization
Your copy
You cannot see the data it holds about you. What you can do

What it is for

Border & ImmigrationRisk Assessment & Triage

This system uses a large language model to help Canada Border Services Agency officers review security screening referrals more efficiently. It reads text data from referrals, flags pre-determined areas of concern, and generates short narrative summaries to support officer decision-making. Officers remain responsible for final screening decisions — the AI is advisory, not determinative.

What it collects and what happens to it

Data taken in

Sensitive personal information
Identifiable data
  • Referral data containing narrative text about individuals under security screening review. This data includes personal information about individuals subject to border security assessment, sourced from STS data (training) and the future Epsilon source system.

Processing

Language Models
  • Open-source large language models (LLMs) are used to identify pre-determined areas of concern in referral text and to generate short narrative summaries for screening officers. The LLM is the only AI component in the project.

What it does

Understanding (Semantic AI)
Human decides
  • The system uses open-source LLM semantic capabilities to identify areas of concern in referral text and retrieve relevant information to support officer review. Officers make all final decisions.
Creating (Generative AI)
Human decides
  • The LLM generates short narrative summaries of referral data to assist screening officers. These summaries are advisory inputs to human decision-making, not binding outputs.

Outputs

A recommendation or prediction
Identifiable data
  • The system outputs flagged areas of concern and short narrative summaries presented to screening officers as advisory inputs. These are recommendations to support human review, not final determinations about individuals.
Generated content
Identifiable data
  • Short narrative summaries of security screening referrals generated by the LLM and presented to officers to assist their review.

Run by

Canada Border Services Agency (CBSA)
  • The Canada Border Services Agency (CBSA) is the federal department deploying and operating the Security Screening Automation Project to support officers reviewing security screening referrals.

Government of Canada AI Register — Security Screening Automation Project

Built by

Government of Canada
  • The Government of Canada developed the Security Screening Automation Project internally, using open-source large language models.

Government of Canada AI Register — Security Screening Automation Project

Kept for

Not stated by the Helpful Places.

Shared with

Available to the accountable organization
  • Outputs generated by the AI system — flagged concerns and narrative summaries — are available to CBSA screening officers and relevant GC employees involved in the review process.
Not available to me
  • Individuals who are the subject of security screening referrals do not have direct access to the AI-generated summaries or flagged concerns produced about them by this system. Access may be possible through formal ATIP (Access to Information and Privacy) requests to the CBSA.

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 UseIndividuals subject to security screening referrals may not be directly informed that an AI system is being used to assist in the review of their file. The Government of Canada's public AI register provides transparency about this system's existence and operation. Further information may be requested through the CBSA.
  • Right to a Human ReviewAll security screening decisions are made by CBSA officers, not by the AI system. The LLM provides only advisory summaries and flagged concerns to support human review. Any individual affected by a screening outcome can seek recourse through CBSA administrative processes.

Risks and safeguards

  • Civil liberties harmThe system operates in a border security and individual security screening context, creating risk of chilling effects on individuals' rights and potential for incorrect flagging of concerns that could influence consequential border decisions.Safeguard: the LLM output is explicitly advisory — screening officers make all final decisions, and the system is designed to support rather than replace human judgment. The system is currently in development, and design choices aim to keep humans in the decision loop.
  • Reputational harmIncorrect identification of areas of concern by the LLM could stigmatize or flag individuals unfairly in a security screening context, with potential downstream reputational consequences.Safeguard: outputs are framed as advisory summaries for officer review, not definitive assessments; human officers validate all outputs before action is taken.