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 Writing and Research Assistant for Government Employees

Employment & Work · Research & Development

What it collects

Operational data
Anonymized data
  • Government of Canada Web Archive (GCWA): 50 TB of data comprising 12.5 billion pages of English, French, and Indigenous text; supplemented by unclassified IM/IT Service Desk articles, Intranet news, and LAC procedural documents. Used as training data, not individual personal records.
About behaviour
Anonymized data
  • Employee prompts and queries submitted to the chatbot at runtime constitute behavioural input data. The register states no personal information is involved; queries are treated as unclassified work interactions.
Run by
Library and Archives Canada (LAC)
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

Employment & WorkResearch & Development

CANChat is a generative AI chatbot that helps Government of Canada employees draft documents, conduct research, and manage knowledge using unclassified government information. It is trained on Canadian data — including 50 terabytes of government web content — and runs on servers located in Canada. Users are informed they are interacting with an AI system, and the tool is intended only for unclassified work.

What it collects and what happens to it

Data taken in

Operational data
Anonymized data
  • Government of Canada Web Archive (GCWA): 50 TB of data comprising 12.5 billion pages of English, French, and Indigenous text; supplemented by unclassified IM/IT Service Desk articles, Intranet news, and LAC procedural documents. Used as training data, not individual personal records.
About behaviour
Anonymized data
  • Employee prompts and queries submitted to the chatbot at runtime constitute behavioural input data. The register states no personal information is involved; queries are treated as unclassified work interactions.

Processing

Language Models
  • A large language model (LLM) trained on Canadian sovereign data, including 50 TB from the Government of Canada Web Archive, hosted in Canadian data centres as part of the Shared Services Canada CANChat initiative.

What it does

Creating (Generative AI)
Human decides
  • The system generates new text in response to employee prompts — drafting documents, composing summaries, and producing research outputs. A government employee reviews and decides whether to use the generated content.
Understanding (Semantic AI)
Human decides
  • The system retrieves and applies meaning from its training corpus — Government of Canada web archives, intranet content, and procedural documents — to ground its generative responses in relevant knowledge.

Outputs

Generated content
Anonymized data
  • The system produces new text — drafted documents, research summaries, knowledge management outputs — in response to employee queries. Outputs are for internal government use on unclassified work only and do not identify individuals.

Run by

Library and Archives Canada (LAC)
  • Library and Archives Canada is the accountable federal department for this AI initiative, providing training data from the Government of Canada Web Archive and overseeing the deployment of CANChat for government employees.

Government of Canada AI Register — CANChat (2526-LAC-BAC-005)

Built by

Shared Services Canada (SSC)
  • Shared Services Canada leads the CANChat initiative, providing the technical infrastructure, hosting in Canadian data centres, and operating the generative AI platform on behalf of the Government of Canada.

Government of Canada AI Register — CANChat (2526-LAC-BAC-005)

Kept for

Not stated by the Helpful Places.

Shared with

Not available to me
  • CANChat is an internal government tool. Members of the public do not have access to query the system or its outputs. Access is restricted to Government of Canada employees for unclassified work.
Available to the accountable organization
  • Outputs and usage data are available to Government of Canada employees and the accountable departments (Library and Archives Canada and Shared Services Canada) for internal operational purposes.

Stored

Stored locally
  • The system is hosted on servers located in Canadian data centres, consistent with the initiative's goals of Canadian data sovereignty and digital independence.
  • Duration: Not specified in register
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 UseGovernment of Canada employees are informed that they are interacting with an AI system when using CANChat. AI use is disclosed to users as confirmed in the official register entry.
  • Right to Algorithmic TransparencyThe Government of Canada has publicly registered this AI system in its Algorithmic Impact Assessment register, disclosing the system's purpose, data sources, training approach, and accountable departments. Employees and the public may consult the register at open.canada.ca.

Risks and safeguards

  • Civil liberties harmThe system is trained on Government of Canada web content, which may include content reflecting historical biases or policy positions that could influence outputs in ways that affect public discourse.Safeguard: The tool is restricted to unclassified internal government work; AI use is disclosed to users; the system does not process personal information; hosting in Canadian data centres limits foreign-jurisdiction access to government data.
  • Societal & cultural harmLLMs can produce plausible-sounding but inaccurate information (hallucination), potentially eroding trust in government communications if outputs are used without verification. Additionally, training on majority-language data may under-represent Indigenous languages and perspectives despite their inclusion in the corpus.Safeguard: The system is designed as a productivity support tool where human employees decide whether to use generated content; the training corpus explicitly includes Indigenous language text from the GCWA.