Algorithmic Impact Assessment
For federal automated administrative-decision systems, complete and publish an Algorithmic Impact Assessment yielding an impact level (I–IV) before production, and scale safeguards to that level.
Canada — Treasury Board · Canada (federal government systems) · Directive (binding on federal institutions)
This page is a plain-English summary written by us, not legal advice — the official text is linked above. Our catalogue was last reviewed 2026-08-15; that is a review of the whole catalogue, not an independent legal verification of this entry.
In force since 2019; fourth-review amendments in force 2025 (existing systems compliant by 24 Jun 2026). Binds federal automated systems supporting administrative decisions — and, in practice, vendors selling to government.
For federal automated administrative-decision systems, complete and publish an Algorithmic Impact Assessment yielding an impact level (I–IV) before production, and scale safeguards to that level.
Give affected individuals notice that a decision is automated, a plain-language explanation, and a means to request review or recourse.
Apply peer review and human-in-the-loop intervention proportionate to the assessed impact level (required at higher levels).
Federal automated administrative-decision systems must meet the fourth-review Directive: a published Algorithmic Impact Assessment, notice/explanation, recourse, and human intervention scaled to impact level.
Action: If you operate or sell federal ADM systems, complete and publish an AIA and wire in the scaled safeguards.
Our agents re-read the official source every few hours and republish this page when it moves. Leave an email and you will hear about it — only when something actually changed.
The same obligation themes side by side — what both demand, what only one does, and which deadline lands first.