evidoria

← Back to browse

Good practice Imported

Canada's Federal AI Register — a First Public Inventory of 409 Government AI Systems

Canada · Ottawa · See the Canada profile · See the Ottawa profile

Evidence: Descriptive / self-reported Top 22% 73/100 · Ask Evidence Copilot about this practice

On 28 November 2025 Canada published its first federal AI Register, an MVP inventory of 409 AI systems across 42 departments and agencies, built on the Treasury Board's 2019 Directive on Automated Decision-Making.

409 systems
AI systems listed in register (28 November 2025)
42 institutions
Federal institutions covered (28 November 2025)

Details

Maturity
Pilot
Promoter
Treasury Board of Canada Secretariat
Period
2019–2025
Keywords
digital government, AI governance, transparency, public administration

Context

Since 2019, Canada's Treasury Board Directive on Automated Decision-Making has required Algorithmic Impact Assessments that score risk before any automated system is deployed.

Objectives

Canada set out to build a public inventory of federal AI use to support oversight, planning and deduplication across government.

Activities

The directive was amended following its fourth formal review, effective June 2025, and the AI Register was launched as a minimum viable product on 28 November 2025 by Treasury Board President Shafqat Ali.

Results

The register lists 409 AI systems, ranging from early exploration to fully operational tools, across 42 federal institutions, recording each system's purpose, description, use and whether it was built in-house or by a vendor.

Conclusions

Because this is an MVP launched only months before this evaluation, independent academic review has already flagged gaps in what departments disclose and omit, so its transparency value will depend on continued completeness checks rather than a one-time publication.

Implementation
Indicative cost
High (€500k–€5M)
Time to results
Long (> 3 years)
Staffing & skills
Treasury Board of Canada Secretariat, 42 federal departments and agencies self-reporting their AI systems

Conditions for success

  • a standing legal instrument (the 2019 Directive on Automated Decision-Making) requiring Algorithmic Impact Assessments before deployment
  • a defined compliance deadline (June 2026) for pre-existing systems to meet updated requirements

Common failure modes

  • as an MVP, independent academic review has already flagged gaps in what departments disclose and omit
  • self-reported entries with no described independent verification of completeness

Commonly funded by

National / regional programmes

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Similar practices you may find useful