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 Communications Drafting for Tax Agency Staff

Inform

What it collects

Operational data
Anonymized data
  • A database containing all published media responses, media lines, and question-and-period responses from the past couple of decades — all previously approved and publicly available information, containing no personal information.
Run by
Canada Revenue Agency (CRA)
Where
No fixed location
Kept
Not stated by the Helpful Places.
Shared with
Accountable organization

What it is for

Inform

This system uses generative AI to help Canada Revenue Agency employees draft communications products — such as media responses, media lines, and question-and-period (QP) responses — more quickly. It draws on a database of already approved and publicly available information. The system does not process personal information, and employees are informed that AI is being used in the drafting process.

What it collects and what happens to it

Data taken in

Operational data
Anonymized data
  • A database containing all published media responses, media lines, and question-and-period responses from the past couple of decades — all previously approved and publicly available information, containing no personal information.

Processing

Language Models
  • Microsoft's generative AI tools, which use large language models, are employed to draft and suggest text for CRA communications products based on a corpus of approved media content.
Search & Retrieval
  • The system retrieves relevant prior approved content — media responses, lines, and QP answers — from the CRA's historical database to inform and ground the AI's generated drafts.

What it does

Creating (Generative AI)
Human decides
  • The system generates draft text for communications products. Human employees review, edit, and approve all outputs before use — the AI produces suggestions, not final published content.
Understanding (Semantic AI)
Human decides
  • The system also performs text and speech analysis to understand queries and retrieve relevant approved content from the communications database, grounding its generative outputs in prior approved material.

Outputs

Generated content
Anonymized data
  • Draft communications products — including media responses, media lines, and QP responses — generated by the AI for review and approval by CRA communications employees. No personal information is included in the output.

Run by

Canada Revenue Agency (CRA)
  • The Canada Revenue Agency (CRA) is the federal department deploying this generative AI tool to support its communications staff in drafting media responses and related products.

AI Register Entry 2526-CRA-ARC-002

Built by

Microsoft
  • Microsoft is the vendor supplying the generative AI tools used by the Canada Revenue Agency for communications drafting.

AI Register Entry 2526-CRA-ARC-002

Kept for

Not stated by the Helpful Places.

Shared with

Available to the accountable organization
  • The AI-generated draft content is available to CRA communications employees (GC employees) who use the tool as part of their workflow. Outputs are reviewed internally before any public release.

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 UseCRA employees who use this system are informed that AI is being used in the drafting process. The AI use is disclosed to users of the tool.
  • Right to Algorithmic TransparencyThe system is disclosed on the Government of Canada's AI Register, providing public information about its purpose, capabilities, data sources, and vendor. Members of the public can access this disclosure through the open government portal.

Risks and safeguards

  • Reputational harmGenerative AI may produce inaccurate, misleading, or off-brand communications drafts that could embarrass the CRA or spread incorrect tax information if published without review.Safeguard: all AI-generated content is reviewed and approved by human CRA communications employees before use; the system is grounded in approved, publicly available source material only.
  • Societal & cultural harmAI-generated government communications could homogenize tone, introduce subtle biases, or erode public trust if errors go undetected.Safeguard: the system is in development, restricted to GC employees as primary users, and all outputs require human approval prior to publication; the system is disclosed publicly via the AI register.