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 Email Triage for Government Mailboxes

Inform · Enforcement

What it collects that can identify you

Sensitive personal information
Identifiable data
  • Incoming emails may contain sender names, email addresses, and personal details submitted by members of the public or other correspondents. The register confirms personal information is involved.

Also collects about behaviour, which is anonymized data.

Run by
Environment and Climate Change Canada (ECCC)
Where
No fixed location
Kept
Retained unspecified
Shared with
Accountable organization
Your copy
You cannot see the data it holds about you. What you can do

What it is for

InformEnforcement

This tool automatically reads incoming emails in a government general-delivery mailbox, tags them with relevant information, and determines whether each email has business value. Emails deemed undesirable — such as spam, phishing attempts, or form letters — are archived rather than acted on. A reporting dashboard lets mailbox owners review email patterns over time. The use of AI in processing these emails is not disclosed to the people who sent them.

What it collects and what happens to it

Data taken in

Sensitive personal information
Identifiable data
  • Incoming emails may contain sender names, email addresses, and personal details submitted by members of the public or other correspondents. The register confirms personal information is involved.
About behaviour
Anonymized data
  • VIP lists are used as a data source, implying tracked correspondence patterns and sender categorization. Email metadata captures behavioural signals such as sending frequency and message patterns that inform spam and form-letter detection.

Processing

Language Models
  • The system uses a language-model-based AI workflow to read and interpret the content of emails, determine business value, and generate tags. The register describes this as 'an AI workflow' that reads every new email and tags it with relevant information.
Classification & Prediction
  • Emails are classified into categories such as spam, phishing, and form letters, and a binary business-value determination is made for each message. Metadata is captured for archived (undesirable) emails.

What it does

Deciding (Analytical AI)
Human decides
  • The AI classifies each incoming email: it tags emails with relevant metadata, assigns a business-value determination (retain vs. archive), and categorizes message type (e.g. spam, phishing, form letter). A dashboard surfaces results for human mailbox owners to review.
Understanding (Semantic AI)
Autonomous
  • The AI reads and understands the content of each email to extract meaning, identify intent (e.g. form letter vs. genuine inquiry), and tag it with relevant information — operating on a scheduled basis without human review of each individual message.

Outputs

A decision about you
Anonymized data
  • The AI makes a binding determination for each email: retain as business-value correspondence or archive/remove as undesirable. Emails flagged as undesirable (spam, phishing, form letters) are automatically archived and removed from active management without human review of each individual case.
About behaviour
Anonymized data
  • The system produces a reporting dashboard that displays the landscape of emails received across time, allowing mailbox owners to see patterns, volumes, and categories of correspondence. Metadata about archived emails is also captured and reported.

Run by

Environment and Climate Change Canada (ECCC)
  • Environment and Climate Change Canada's Digital Services Branch built and operates this correspondence monitoring tool to manage the department's general-delivery mailbox.

Government of Canada AI Register — 2526-ECCC-005

Built by

Government of Canada
  • The Government of Canada developed this AI system internally. The register lists 'Government of Canada' as the developer, indicating in-house development rather than a third-party vendor.

Government of Canada AI Register — 2526-ECCC-005

Kept for

Retained unspecified
  • Undesirable emails (spam, phishing, form letters) are archived along with captured metadata. The register does not specify a retention period for archived emails or dashboard data.
  • Duration: unspecified

Shared with

Available to the accountable organization
  • Outputs — including email classifications, tags, archived metadata, and dashboard reports — are available to GC employees (mailbox owners) within Environment and Climate Change Canada's Digital Services Branch.
Not available to me
  • Members of the public who send emails to the general-delivery mailbox are not informed that their correspondence is being processed by AI, and have no access to how their email was classified or whether it was archived. The register explicitly states that AI use is not disclosed to users.

Stored

Stored locally
  • The system is operated by a Canadian federal government department. Storage is presumed to be within Canadian jurisdiction, consistent with Government of Canada data residency requirements, though the register does not specify storage location or cloud provider.
  • Duration: unspecified
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 UseMembers of the public who email the general-delivery mailbox are currently not informed that their correspondence is processed by an AI system. The register explicitly states that AI use is not disclosed to users. Until a disclosure mechanism is implemented, this right is not fulfilled. Individuals who wish to raise concerns about AI use may contact Environment and Climate Change Canada at ec.general-generale.ec@canada.ca.
  • Right to Algorithmic TransparencyThe general logic of this system — that incoming emails are automatically classified and potentially archived by AI — is disclosed through the Government of Canada's AI Register. However, this information is not proactively surfaced to members of the public who send emails to the monitored mailbox.

Risks and safeguards

  • Civil liberties harmMembers of the public submit correspondence to a government mailbox without being informed their emails are processed by AI. This lack of disclosure could undermine the right to know how personal information is handled by public institutions. Automated archiving of emails as 'form letters' or 'spam' could suppress legitimate constituent communication.Safeguard: The system is described as a proof-of-concept still in development; human mailbox owners retain dashboard oversight. However, no disclosure mechanism to email senders is documented, representing a gap.
  • Reputational harmLegitimate emails from members of the public or stakeholders could be misclassified as spam or form letters and archived, causing their communications to go unacknowledged or unprocessed by the department. This could harm the sender's ability to engage with government and damage institutional credibility.Safeguard: The metadata is captured for all archived emails and a reporting dashboard is available to mailbox owners for review; however, no documented appeal or correction mechanism is described for senders whose emails are incorrectly archived.