Canada Builds AI Framework to Navigate Patchwork Regulations

Researchers at an unspecified institution present SCITUS (Systematic Canadian Integration for Trustworthy and Unified Standards), a framework adapting the NIST AI Risk Management Framework to both federal and provincial AI regulations in Canada. The framework addresses a fragmented legal landscape following the collapse of Bill C-27 and the Canadian government’s shift toward targeted AI legislation.

SCITUS integrates seven trustworthy-AI characteristics and four governance functions, alongside a methodology for mapping compliance across multiple jurisdictions. The framework’s control catalog has expanded from 31 controls in June 2025 to 57 in July 2026, reflecting evolving regulations and emerging threats, including findings related to generative-AI training data and agentic-AI risks.

Researchers demonstrated SCITUS’s applicability in government, healthcare, and the private sector. They suggest a systematic approach to adapting global AI frameworks offers benefits over attempting compliance on a jurisdiction-by-jurisdiction basis and could serve as a model for other federal systems.