Canada
1) Mandatory Compliance for Legacy Systems (January 5, 2026)
The Treasury Board of Canada Secretariat reaffirmed that all automated decision-making systems developed prior to June 2025 must achieve full compliance with the Directive on Automated Decision-Making by the 2024–2026 window. This requires mandatory Algorithmic Impact Assessments (AIA) and the integration of human intervention capabilities into system architectures.
Directive on Automated Decision-Making
How it applies to AI in Healthcare:
This applies to AI used in healthcare procurement and federal health services. It mandates that legacy clinical-support tools undergo impact assessments to ensure safety and transparency, ensuring they do not operate without human-in-the-loop safeguards.
2) Health Canada 2025-26 Departmental Plan (January 7, 2026)
Health Canada identified the modernization of health care through digital tools as a primary pillar for the fiscal year. This includes the implementation of the Pan-Canadian Interoperability Roadmap to establish technical standards for secure, longitudinal patient data access across provincial borders.
Health Canada 2025-26 Departmental Plan
How it applies to AI in Healthcare:
Standardized data interoperability is the foundation for AI training and deployment in Canada. This update ensures that AI models can access high-quality, cross-jurisdictional data while adhering to a unified national framework for digital health tools.
3) AIDA and High-Impact AI Classification (January 9, 2026)
As Bill C-27 (AIDA) progresses through the Senate, new session notes emphasize the classification of AI in healthcare as “High-Impact.” This status will trigger rigorous safety, fairness, and accountability audits once the Act is fully proclaimed.
C-27 (44-1) – LEGISinfo – Bill Status Update
How it applies to AI in Healthcare:
AI tools used for medical triage, biometric identification, or patient diagnostics will be subject to the highest level of regulatory scrutiny in Canada, requiring developers to provide extensive documentation on bias mitigation and algorithmic validity.
United States
1) FDA Agentic AI Integration (January 2026)
Following the December 2025 announcement, the FDA entered January 2026 by hosting its Scientific Computing Day, demonstrating “Agentic AI” solutions. These systems, which can plan and execute multi-step actions with human oversight, are being integrated into pre-market reviews and post-market surveillance.
FDA Expands Artificial Intelligence Capabilities with Agentic AI Deployment
How it applies to AI in Healthcare:
Regulators are now using AI themselves to speed up reviews; developers must ensure their AI submissions are compatible with “agentic” automated oversight.
2) TEMPO Pilot Launch (January 2, 2026)
The FDA began collecting statements of interest for the Technology-Enabled Meaningful Patient Outcomes (TEMPO) pilot. This program allows for enforcement discretion for certain digital health devices (including AI-enabled software) that can demonstrate improved patient outcomes through real-world data collection.
Digital Health Center of Excellence: TEMPO Pilot
How it applies to AI in Healthcare:
Offers a regulatory shortcut for AI tools that prioritize real-world patient outcomes over traditional, static pre-market clinical trials.
European Union & United Kingdom
1) UK AI Growth Lab Consultation Closure (January 7, 2026)
The UK Government closed its “Call for Evidence” for the AI Growth Lab. This lab is designed to allow the temporary modification or “disapplication” of certain regulatory requirements to foster AI innovation, with the MHRA looking at the “AI Airlock” for medical devices.
AI Growth Lab: Call for Evidence
How it applies to AI in Healthcare:
Signals a transition to a “sandbox” environment where high-potential AI health tools can be tested in the real world with reduced regulatory friction.
Rest of the world
1) UN/WHO White Paper on AI Healthcare Governance (January 1, 2026)
A new UN-backed white paper was released, categorizing global governance into “Market-Driven,” “Rule-Driven,” “Value-Oriented,” and “Agile.” The paper calls for “Differentiated Regulation” and “Full-Process Coverage” for AI in healthcare.
White Paper on AI Healthcare Governance
How it applies to AI in Healthcare:
Provides a global taxonomy for AI regulation, encouraging developers to prepare for “full-process” oversight from data ingestion to clinical impact.
2) South Asia MedTech Expansion (January 9, 2026)
The World Intellectual Property Organization (WIPO) and the International Telecommunication Union (ITU) highlighted new AI-driven screening systems in Sri Lanka and Bhutan, focusing on integrating AI datasets into Intellectual Property (IP) frameworks.
Transforming Eye Care in Sri Lanka
How it applies to AI in Healthcare:
Highlights the growing importance of protecting clinical datasets as IP, especially in rural or emerging health tech markets.
Cross-Cutting Themes
-
Retrospective Accountability: Regulators are no longer grandfathering in older AI. Canada’s June 2026 deadline for legacy systems mirrors a global trend of “cleaning up” existing clinical algorithms to meet modern transparency standards.
-
The Rise of “Agentic” & Multi-Agent Oversight: There is a transition from monolithic AI to “Agentic” systems that can execute multi-step actions. Both the FDA and Canadian health leaders are moving toward architectures where specialized AI agents (e.g., one for lab trends, one for drug conflicts) provide built-in checks and balances.
-
Interoperability as a Regulatory Gatekeeper: The common thread across the Canada and US updates is that data access is now tied to compliance. Without adhering to the Pan-Canadian Interoperability Roadmap or the FDA’s RWE standards, AI tools will lack the data “fuel” required for certification.
Immediate, concrete checklist for health organisations & vendors
-
Inventory Legacy AI: Audit all clinical and administrative AI systems deployed before 2025. Ensure they have a completed Algorithmic Impact Assessment (AIA) and a documented “human-in-the-loop” override before the June 2026 deadline.
-
Validate Interoperability Standards: Ensure your AI’s data ingestion layer is compatible with the Pan-Canadian Interoperability Roadmap. Focus on FHIR standards to enable cross-provincial data exchange.
-
Prepare for High-Impact Audits: If your tool assists in triage or diagnosis, begin a “Bias and Fairness” audit immediately. Document the demographic diversity of training sets to meet the “High-Impact” requirements expected in the 2026 regulatory cycle.
-
Explore Sandbox Eligibility: Evaluate if current AI products qualify for the FDA TEMPO pilot or the UK AI Growth Lab to bypass traditional pre-market hurdles in exchange for live monitoring.
Sources
- Directive on Automated Decision-Making (2025-06-24)
https://www.tbs-sct.canada.ca/pol/doc-eng.aspx?id=32592
- Health Canada 2025-26 Departmental Plan (2025-06-17)
https://www.canada.ca/en/health-canada/corporate/transparency/corporate-management-reporting/report-plans-priorities/2025-2026-departmental-plan.html
- C-27 (44-1) – LEGISinfo – Bill Status Update (2026-01-06)
https://www.parl.ca/legisinfo/en/bill/44-1/c-27
- FDA Expands AI Capabilities with Agentic AI (2025-12-01)
https://www.fda.gov/news-events/press-announcements/fda-expands-artificial-intelligence-capabilities-agentic-ai-deployment
- Digital Health Center of Excellence: TEMPO Pilot (2025-12-09)
https://www.fda.gov/medical-devices/digital-health-center-excellence
- UK AI Growth Lab: Call for Evidence (2026-01-07)
https://www.gov.uk/government/calls-for-evidence/ai-growth-lab/ai-growth-lab
- White Paper on AI Healthcare Governance (2026-01-01)
https://unpan.un.org/resources/white-paper-ai-healthcare-governance
- Transforming Eye Care in Sri Lanka (2026-01-09)
https://unsdg.un.org/latest/stories/transforming-eye-care-sri-lanka-dhanushi%E2%80%99s-journey-ophtha-innovations
