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-Powered Document Search and Analysis for Diplomacy

Inform · Planning & Decision-making · Research & Development

What it collects

Operational data
Anonymized data
  • Unclassified internal GAC documents including SITREPs (situation reports), readouts, and statements. These are institutional records about events, meetings, and positions — not about individual citizens.
About a place
Anonymized data
  • External news via Factiva, covering international events and locations relevant to Canadian foreign policy. The system monitors documents for mentions of key locations.
Run by
Global Affairs Canada (GAC)
Where
No fixed location
Kept
Retained not specified in the register
Shared with
Accountable organization
Your copy
You cannot see the data it holds about you. What you can do

What it is for

InformPlanning & Decision-makingResearch & Development

This system helps Global Affairs Canada officials search, summarize, and monitor large volumes of government and news documents to support foreign-policy work. It uses AI to identify key topics, trends, and mentions of people, places, and organizations across documents like situation reports and readouts. The system processes unclassified documents and does not handle personal information. Users are not currently informed that AI is in use.

What it collects and what happens to it

Data taken in

Operational data
Anonymized data
  • Unclassified internal GAC documents including SITREPs (situation reports), readouts, and statements. These are institutional records about events, meetings, and positions — not about individual citizens.
About a place
Anonymized data
  • External news via Factiva, covering international events and locations relevant to Canadian foreign policy. The system monitors documents for mentions of key locations.

Processing

Language Models
  • GPT-3 is used for embedded document summarization. A lightweight LLM proxy supports flexible model interfacing within the system architecture.
Search & Retrieval
  • Azure Cognitive Search with a sophisticated indexing platform enables users to search across large volumes of documents. Custom NLP skillsets extract GAC-specific entities to improve retrieval precision.
Affect & Emotion Analysis
  • Machine learning is applied for sentiment analysis of document content, helping officials gauge the tone and positioning reflected in diplomatic and news documents.

What it does

Understanding (Semantic AI)
Human decides
  • Uses Azure Cognitive Search and custom NLP skillsets to index documents, recognize named entities (people, locations, organizations), and match user queries to relevant documents. Officials review results and make decisions.
Deciding (Analytical AI)
Human decides
  • Applies machine learning for sentiment analysis and named entity recognition to classify and score document content. Results inform officials who decide how to act on the analysis.
Creating (Generative AI)
Human decides
  • Uses GPT-3 for embedded document summarization, generating concise summaries of documents for review by officials who then use those summaries to inform their work.

Outputs

A recommendation or prediction
Anonymized data
  • Search results, ranked document lists, identified emerging trends, entity mentions, and sentiment scores. These are advisory outputs reviewed by officials — the system does not make binding decisions.
Generated content
Anonymized data
  • AI-generated document summaries produced by GPT-3. These summaries are embedded within the tool for officials to review and do not name or identify individual citizens.

Run by

Global Affairs Canada (GAC)
  • The federal department responsible for Canada's international relations, trade, and consular services. Global Affairs Canada developed and operates Document Cracker to support officials in analyzing diplomatic documents.

GC AI Register — Document Cracker

Built by

Government of Canada
  • The system was developed by the Government of Canada. It uses Microsoft Azure services (Azure Cognitive Search, Power Automate, Logic Apps), Databricks, and OpenAI GPT-3 as underlying technology components.

GC AI Register — Document Cracker

Kept for

Retained not specified in the register
  • The register does not specify a retention period for documents or outputs processed by the system. Standard Government of Canada information management policies would apply.
  • Duration: not specified in the register

Shared with

Available to the accountable organization
  • Outputs — search results, entity mentions, summaries, sentiment scores, and Power BI visualizations — are available to GC employees (primarily Global Affairs Canada officials) who use the system.
Not available to me
  • The system is an internal government tool. Outputs are not available to members of the public. The register also notes that AI use is not disclosed to users of affected services.

Stored

Stored on 3rd Party Cloud
  • The system runs in a secure cloud environment using Microsoft Azure services (Azure Cognitive Search, Power Automate, Logic Apps) and Databricks, which are third-party cloud platforms.
  • Duration: not specified in the 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 Algorithmic TransparencyThe Government of Canada AI Register entry for Document Cracker is publicly available at open.canada.ca and discloses that AI is in use. However, the register notes that AI use is not currently disclosed to users of the affected services. Citizens and officials seeking more information may consult the AI Register entry at the source URL.
  • Right to Be Informed of AI UseThe register explicitly states that AI use is not disclosed to users of the affected services. This means individuals interacting with outputs influenced by this system may not be informed that AI was involved. This is a transparency gap acknowledged in the official register.

Risks and safeguards

  • Reputational harmNamed entity recognition and sentiment analysis on diplomatic documents may misidentify individuals or mischaracterize the tone of communications, potentially leading to incorrect assessments of people or countries.Safeguard: The system produces advisory outputs only; officials review results before acting. The system runs in a secure cloud environment and processes only unclassified documents.
  • Civil liberties harmAutomated monitoring of documents for mentions of key people, locations, and organizations could — if scope expanded — raise surveillance concerns. The register states no personal information is involved and the system is limited to unclassified internal and public news documents.Safeguard: Scope is explicitly limited to unclassified documents; the system does not handle personal information according to the register entry.