Case 1: Managing the Machine — Innovare AI Analysis

As you read — hold this question

If your AI tells a customer something that isn't true, whose words are they — the company's, or the bot's?

$812

Air Canada's website chatbot invented a bereavement-refund policy. A grieving passenger relied on it. A tribunal ruled the airline liable for CA$812 — and rejected the claim that the bot was “a separate legal entity.”

Module 10 asks how organisations actually manage AI. The examinable spine is the seven aspects of managing AI, Drucker's definition of management, and the knowledge/data discipline underneath. Air Canada (Feb 2024) is what happens when those aspects are skipped.

The examinable core

The textbook versionQuiz: the 7 aspects + Drucker

Management, the 7 aspects, and knowledge

Management (Drucker 2008): “the activities of communicating, leading, coordinating and controlling the tasks of others,” with decision-making the fundamental managerial activity. AI is described as the new frontier of that decision-making.

The 7 aspects of managing AI (IBM/ITDart/NAA): 1) Strategic Integration, 2) Data Management, 3) Risk Management, 4) Governance & Ethics, 5) Talent & Training, 6) Monitoring & Evaluation, 7) Continuous Improvement.

Knowledge management shifts from manual→instant retrieval, passive→proactive sharing, periodic→real-time alignment, information-based→AI-driven decisions. Best practices for information/data: governance framework; data quality & integration; lifecycle management; compliance & security.

Carry this

Learn the seven aspects in order. Air Canada failed 3, 4 and 6 — Risk, Governance and Monitoring — and paid for it.

Map it to the failure

The rest of the story

What the frameworks did and didn't anticipate

Aspect	What broke at Air Canada
Risk Management	No assessment of what the bot might invent.
Governance & Ethics	No rule for “the AI stated a policy that doesn't exist.”
Monitoring & Evaluation	Nobody was watching the outputs for hallucinated policy.
Data Management	The bot answered from statistical plausibility, not the real fare rules.

The honest note

The course's aspect list comes from vendor/government sources (IBM, ITDart, NAA). They are useful checklists but descriptive, not predictive — none anticipated the hallucinated-policy failure mode. Frameworks tell you what to manage; they don't tell you what will go wrong.

The lesson

The other side

Your AI's words are your words

The tribunal's principle is the whole case: a company is responsible for what its AI says, exactly as for any other part of its website. The most important managerial capability here is not sensing or seizing — it is knowing when the model is making things up, and keeping a human accountable for it.

The other side

Governance is not aspect 4 of 7 — it is the one that decides whether the other six are safe to use.

Quick recall — without looking back

Test yourself on this case

Question 1 of 3

List the seven aspects of managing AI.

Strategic Integration; Data Management; Risk Management; Governance & Ethics; Talent & Training; Monitoring & Evaluation; Continuous Improvement.

Question 2 of 3

What is Drucker's definition of management, and what is the “fundamental activity”?

“The activities of communicating, leading, coordinating and controlling the tasks of others,” with decision-making as the fundamental managerial activity.

Question 3 of 3

What principle did the Air Canada tribunal establish?

A company is liable for what its AI chatbot tells customers — it is not a separate legal entity. Governance, not the model, was the failure.

Module 10

← Module Overview

Case 2: Sense, Seize, Reconfigure →

Module 10 Videos

Module 10 · Long Form

Sources

Module content: BUSN9049 Module 10 — managing AI, the 7 aspects, knowledge/data management (deck + videos). Flinders University, 2026.
Case: Moffatt v. Air Canada, 2024 BCCRT 149 (Civil Resolution Tribunal, Feb 2024).
Frameworks: IBM (2024) AI governance; National Archives of Australia, information management; ITDart (2024).