Directive on Automated Decision-Making (Canada)

• Authority: Government of Canada
• Jurisdiction: Other countries
• Collections: Miscellaneous documents
• Most recent activity: Enacted
• Most recent activity date: 2023-04-25

Risk subdomain coverage

Good coverage — Comprehensive governance measures or mitigations with clear procedures relating to the risk subdomain are explicitly described.

• Unfair discrimination and misrepresentation
• Unequal performance across groups
• Compromise of privacy
• AI system security vulnerabilities
• Governance failure
• Lack of capability or robustness
• Lack of transparency or interpretability

Minimal coverage — Brief mention but no specific mitigations, recommendations or governance measures relating to the risk subdomain are described.

• False or misleading information
• Overreliance and unsafe use
• Loss of human agency and autonomy

All other risk subdomains had no coverage in this document.

Risk Coverage Summary

The document has good coverage of approximately 8-10 subdomains, with strong focus on unfair discrimination (1.1, 1.3), privacy compromise (2.1), AI system security (2.2), false information (3.1), lack of transparency (7.4), lack of robustness (7.3), and governance failure (6.5). Coverage is concentrated in discrimination/bias mitigation, transparency, quality assurance, and governance domains.

Sector coverage

Good coverage — Comprehensive governance measures or mitigations with clear procedures relating to the sector are explicitly described.

• Public Administration (excl. National Security)

All other sectors had no coverage in this document.

Sectors Summary

This directive primarily governs the Public Administration sector, as it applies to all federal institutions using automated decision systems to make administrative decisions about clients. It does not regulate private sector activities but rather internal government operations across all federal departments and agencies (excluding Agents of Parliament).

Legislative Status

Hard Law

Legislative Status Summary

This is a binding directive issued under the authority of the Financial Administration Act with mandatory compliance requirements, specific enforcement mechanisms, and consequences for non-compliance including measures allowed by the Financial Administration Act.

Actors

Proposer

• Entity Types: AI_Governance_Actor
• Summary: The directive is issued by the Treasury Board of Canada Secretariat under the authority of the Policy on Service and Digital, with the Chief Information Officer of Canada responsible for developing and maintaining the Algorithmic Impact Assessment and providing government-wide guidance.

Target

• Entity Types: AI_Deployer, AI_Governance_Actor
• Summary: The directive applies to all federal institutions that use automated decision systems to make administrative decisions about clients, with specific exclusions for Agents of Parliament. The Assistant Deputy Minister responsible for the program using the automated decision system is the primary target for compliance.

Enforcer

• Entity Types: AI_Governance_Actor
• Summary: The Treasury Board has authority to determine appropriate measures for non-compliance under the Financial Administration Act, and the Treasury Board of Canada Secretariat oversees implementation and compliance through the Chief Information Officer of Canada.

Monitor

• Entity Types: AI_Governance_Actor
• Summary: The Chief Information Officer of Canada is responsible for government-wide monitoring and engagement, while individual institutions are required to monitor their own automated decision systems on a scheduled basis and publish effectiveness information. The directive itself is reviewed every two years.

AI Lifecycle Stages

Lifecycle stages covered by the documents, as defined in the OECD Framework for the Classification of AI Systems

• Plan and Design
• Collect and Process Data
• Build and Use Model
• Verify and Validate
• Deploy
• Operate and Monitor

AI Lifecycle Summary

The document comprehensively covers all stages of the AI lifecycle, with particularly strong emphasis on Plan and Design (through Algorithmic Impact Assessments and Gender-based Analysis Plus), Verify and Validate (through bias testing, peer review, and quality assurance), Deploy (through transparency and approval requirements), and Operate and Monitor (through ongoing monitoring, reporting, and recourse mechanisms). Data collection and processing requirements are also explicitly addressed through data quality and governance provisions.

Technical Scope

Terms relating to technical scope that are mentioned or implicitly referenced in the document

• AI Models
• AI Systems
• Open Weight / Open Source

Technical Scope Summary

The document explicitly covers automated decision systems, which are defined as any system, tool, or statistical model used to make administrative decisions. It does not specifically mention AI models, AI systems, frontier AI, general purpose AI, task-specific AI, foundation models, generative AI, predictive AI, open-weight models, or compute thresholds. The focus is on automated decision-making systems in the context of government administrative decisions.