← Global AI Laws & Governance Research Library
Voluntary Code of Conduct on the Responsible Development and Management of Advanced Generative AI Systems, Canada
Prepared by Shubhank Suman · Last reviewed
CURRENT / VOLUNTARY
- Legal character
- Voluntary AI governance code
- Region
- North America
- Announced
- 27 September 2023
What it covers
Canada's Code applies to organisations developing or managing advanced generative AI systems with broad capabilities. Narrowly designed low-risk tools generally sit outside its intended scope. It provides a voluntary governance baseline organised around Safety, Accountability, Transparency, Fairness and Equity, Human Oversight and Monitoring, and Validity and Robustness. The implementation guide identifies 18 operational measures.
Why it matters
It provides a practical voluntary governance layer for advanced general-purpose generative systems.
Refer to it when
Use it for advanced GenAI development or management, governance responsibility, testing, monitoring, transparency, fairness, human oversight, robustness and security.
AI, privacy and cybersecurity relevance
- AI: Very High
- Privacy: High
- Cyber: Very High
AI
VERY HIGH RELEVANCE
The Code establishes six governance principles and 18 implementation measures for advanced generative AI developers and managers.
Privacy
HIGH INDIRECT RELEVANCE
The Code does not replace PIPEDA, provincial privacy statutes, sector rules or contracts. Privacy issues include training and fine-tuning data, prompts, sensitive information, outputs, retention and third-party AI services.
Cybersecurity
VERY HIGH RELEVANCE
Safety, robustness, monitoring and accountability directly concern adversarial attacks, poisoning, manipulation, prompt attacks, unauthorised access, misuse and incident management.
Applicability
Voluntary commitments for organisations developing or managing advanced generative AI systems within the Code's scope.
Current validity
CURRENT VOLUNTARY CODE. This instrument is Canadian, not South Korean.
Read first
Safety → Accountability → Transparency → Fairness & Equity → Human Oversight & Monitoring → Validity & Robustness → 18 implementation measures.
Provisions / sections covered
Safety
Identify foreseeable misuse, harmful outputs, system limitations and security threats, and implement mitigation.
Accountability
Assign responsibility and establish internal risk-management structures.
Transparency
Provide appropriate system information and help users understand when AI is being used.
Fairness & Equity
Identify and mitigate discriminatory or unfair outcomes.
Human Oversight & Monitoring
Monitor performance and behaviour, detect incidents, maintain human oversight and respond to material issues.
Validity & Robustness
Test and monitor reliability, robustness, security and resistance to malicious attacks.
Subject-matter tags
- Retention
- Enforcement and penalties
- Surveillance
- AI governance
- Generative AI
- Transparency
- Human oversight
- Risk management
- AI cybersecurity
- Accountability
Subject-matter tags describe what the indexed text discusses. They are not legal conclusions.
Related instruments in this library
- Artificial Intelligence and Data Act (AIDA), CanadaBill · HistoricalSame jurisdiction; assess applicability separately.
- Digital Charter Implementation Act, 2022: Bill C-27: CanadaBill · HistoricalSame jurisdiction; assess applicability separately.
- Operationalizing the Gen-AI Voluntary Code of Conduct for Canadian Critical Infrastructure EntitiesGuidance · Voluntary / non-bindingSame jurisdiction; assess applicability separately.
- Personal Information Protection and Electronic Documents Act: PIPEDA: CanadaLaw · CurrentSame jurisdiction; assess applicability separately.
- Canada's Anti-Spam Legislation (CASL), S.C. 2010, c. 23Law · Current (phased commencement)Same jurisdiction; assess applicability separately.
Document page in the Knowledge Base - full text or PDF where it is available for this record.
This library locates primary and official sources. It does not interpret the law and it is not legal advice.