Audience:
Executives, managers and decision owners in government (including defence and security) and the private sector
Duration:
0.5 Day
Delivery:
Virtual (1 × half day) or in-person (1 × half day). Delivered live-virtual across Canada and the US, and in-person in Ottawa–Gatineau (National Capital Region) and Toronto. On-site delivery available for Government of Canada and private-sector teams.
Next public sessions (live-virtual, 1:00-4:00pm ET , $380/person): Dec 7 2026 · Feb 19, 2027 — book on our training page.
Course description:
This course explains how AI supports, augments, and in some cases automates decision‑making, and what this means for responsibility, oversight, and mission assurance
Course overview:
The anatomy of a decision: a brief, more philosophical look at what a decision is, separating what exists (structure) from what it’s for (purpose). It covers why asking the right question comes before any data or model.
Taking a decision apart: the steps inside a decision, and which of them AI can inform, recommend or decide.
Confidence, thresholds and errors: why an accurate system can still be wrong most of the time it raises a flag, and how to set the threshold where a system acts or hands over to a person, based on what each kind of error costs.
People in the loop: designing human oversight that works rather than rubber-stamps, and keeping authority and accountability clear when AI contributes to a decision.
Algorithmic Impact Assessment in depth: how Canada’s Algorithmic Impact Assessment weighs risk against mitigation, how a system lands at an impact level, and what each level asks of the organization that deploys it. It also shows how the same thinking carries over to the private sector.
After go-live: monitoring, model drift, and knowing when an automated decision needs to come back to a person.
By the end of this course, participants will be able to:
Explain what a decision is, and why framing the question matters before choosing data or AI
Break a decision into its steps and identify which ones AI should inform, recommend or decide
Explain how confidence thresholds and the cost of errors determine when a system acts on its own and when it refers to a person
Design human oversight that keeps authority and accountability clear and avoids rubber-stamping
Interpret how an Algorithmic Impact Assessment determines a system’s impact level and the oversight it requires
Plan the monitoring and review an AI-assisted decision needs once it’s in use
Before you come: no technical background is needed. AI Fundamentals for Professionals is useful preparation but not required.
Interactive Data Tools for This Course
These tools are free to use anytime — browse them before the course as a preview, work through them live during a session, or come back to them afterward as a refresher.