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

Back to your search

AI-Assisted Publication Standards Check for Census Documents

Enforcement

What it collects

Operational data
Anonymized data
  • Publication drafts for census outputs, written in French or English, submitted for standards compliance checking before dissemination. These are institutional documents, not personal data.
Run by
Statistics Canada (StatCan)
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

Enforcement

This tool uses a large language model to automatically review Statistics Canada census publication drafts and flag passages that may not comply with official dissemination standards. It is used by Government of Canada employees to ensure consistent adherence to standards before public release, allowing human reviewers to focus on content quality. The system does not process personal information. AI use is disclosed to users of the tool.

What it collects and what happens to it

Data taken in

Operational data
Anonymized data
  • Publication drafts for census outputs, written in French or English, submitted for standards compliance checking before dissemination. These are institutional documents, not personal data.

Processing

Language Models
  • A large language model (LLM) processes French and English publication draft text, performing text and image analysis to identify passages that do not meet Statistics Canada dissemination standards.

What it does

Deciding (Analytical AI)
Human decides
  • The system classifies and flags sections of publication drafts as compliant or non-compliant with dissemination standards, returning a set of flagged issues. Human reviewers then decide whether to act on each flag.
Understanding (Semantic AI)
Human decides
  • The LLM reads and understands French and English publication draft text, grounding its analysis in the knowledge of StatCan standards to identify relevant passages for review.

Outputs

A recommendation or prediction
Anonymized data
  • A set of flagged issues identifying passages in publication drafts that may not comply with StatCan standards, presented to human reviewers for their decision. The system recommends but does not decide; intervention is human.

Run by

Statistics Canada (StatCan)
  • Statistics Canada is the federal department responsible for producing statistics to help Canadians better understand their country. It is the deploying organization accountable for this AI tool, which is used by its employees to check census publication drafts for compliance with dissemination standards.

Government of Canada AI Register — 2526-StatCan-023

Built by

Government of Canada
  • The system was developed by the Government of Canada. As developer and deployer are both within the federal government, this entry reflects the development-side accountability.

Government of Canada AI Register — 2526-StatCan-023

Kept for

Not stated by the Helpful Places.

Shared with

Not available to me
  • The system is used by Government of Canada employees (internal tool). Members of the public do not interact with or have access to the system or its outputs directly.
Available to the accountable organization
  • Output flags and reviewed results are available to Statistics Canada employees involved in the publication review workflow.

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 UseAI use is disclosed to users of the tool. Government of Canada employees using this system are informed that AI is being used to assist in the publication standards review process.
  • Right to a Human ReviewThe system is designed to flag issues for human intervention. All determinations about publication compliance are made by human reviewers, not the AI system alone. Human oversight is a core design principle of the tool.

Risks and safeguards

  • Reputational harmThe LLM may incorrectly flag compliant passages or miss non-compliant ones, potentially affecting the reputation of publications or their authors if errors propagate.Safeguard: The system is designed to flag issues for human intervention, not to make final decisions. Human reviewers retain authority over all publication determinations, limiting the risk of unchecked automated errors.
  • Societal & cultural harmIf the system systematically misinterprets standards for French-language documents, it could create inequitable review burdens or erode trust in bilingual publication quality.Safeguard: The system is described as supporting both French and English text analysis. Human reviewers remain the decision-makers, preserving bilingual professional oversight.