Behavioral Science of AI Governance
-
78%
of executives doubt they can pass an AI audit. (Grant Thornton)
-
60 minutes
of course content spread across 6 modules and 11 videos. Learn at your own pace!
-
5
downloadable files, worksheets, and templates that allow you to assess AI governance in your organization.
-
>97%
of previous participants in masterclasses, workshops, and executive education recommend the instructor to others.
Why this course exists.
By the time an AI initiative reaches a board, the individual‑level resistance and team‑level process drift have usually already been handled—or already gone wrong—and what remains is a higher‑level failure: approval without oversight. Boards keep greenlighting major AI investments even though nearly half haven’t set governance expectations, almost as many haven’t built AI risk into ongoing oversight, 78% of executives doubt they could pass an audit within 90 days, and 66% of boards still lack working AI knowledge, meaning they can’t meaningfully challenge what they approve. Even when a governance framework exists on paper, fewer than a quarter have implemented real controls. These aren’t just knowledge gaps but the same behavioral patterns seen earlier, now one layer up: escalation of commitment when the approvers are also the evaluators, and diffused responsibility when oversight sits with a committee where everyone is responsible and no one is. This course treats governance failure as a behavioral problem, giving leaders and boards a way to diagnose which failure mode is operating in their approval and oversight process and design around it.
What you’ll leave with.
-
A documented AI approval decision from your own organization, diagnosed against the evidence-vs-confidence framework.
-
A set of kill criteria for one active AI initiative, plus an honest read on whether current support for it is evidence or sunk cost.
-
A redesigned reporting path for one escalation point, with a reduced cost to surfacing bad news to leadership.
-
An Ownership Map for one AI governance decision, naming who's actually accountable rather than who's in the room.
-
A Governance Audit that combines the four into a single diagnosis-and-action plan—and connects back to your Adoption diagnosis and Drift audit if you've completed those courses.
-
-
-
Why do committees fund confident, well-delivered pitches over the evidence in front of them? This video diagnoses approval bias—the deference to seniority and pilot-stage optimism that lets conviction substitute for proof in AI funding decisions.
-
Now that you can see approval bias, this video builds a process that catches it—one that forces evidence to be weighed before conviction, on every AI pitch that reaches a governance committee. You'll walk away with a repeatable check you can apply to your next approval decision.
-
-
-
Six months after a confident pitch gets approved, the metrics turn—so why does the committee approve one more quarter instead of killing it? This video diagnoses escalation of commitment: the sunk-cost thinking and self-evaluation bias that keep failing AI initiatives funded long after the data says stop.
-
This video builds three structural defenses against escalation of commitment: kill criteria set before launch, a fresh-eyes test that strips out sunk cost, and independent review by someone who wasn't part of the original approval. None of them depend on anyone being more disciplined in the moment it matters most.
-
-
-
The people closest to a failing AI initiative often see the problem first—so why doesn't leadership hear about it? This video diagnoses the incentive gap: the well-documented reluctance to deliver bad news, and the cost-versus-payoff calculation that keeps concerns from ever reaching the top.
-
This video designs a reporting path that actually works: a low-cost channel for raising concerns, a closed loop that proves speaking up leads somewhere, and a leadership habit that rewards the messenger instead of merely tolerating them. The goal isn't asking people to be braver—it's changing what's rational for them to do.
-
-
-
A committee can discuss an AI decision thoroughly and still produce one that nobody individually owns. This video diagnoses diffused ownership—groupthink, diffusion of responsibility, and the "everyone's accountable, so no one is" pattern behind governance failures like Boeing's missing safety oversight.
-
his video builds three fixes for diffused ownership: silent written input before discussion, a named owner attached to every governance decision, and a standing agenda item that can't quietly disappear. Together, they make sure a decision was actually examined by individuals—and owned by one of them.
-
-
-
The course's closing module shows how all four failure modes compound on each other, using a well-established model from safety science to explain why. You'll see how a person-level bias, a team-level blind spot, and an organizational gap can align on the same decision—and why that alignment, not any single failure, is what actually breaks governance.
-
This final video builds the Governance Audit: four questions, pulled from every module in this course, that check every layer of the stack in about five minutes. Run it before you approve an AI initiative, and run it again at every review after.
-
Your instructor.
Eugene Chan, PhD
Eugene (PhD Toronto; MA Chicago; AB Michigan) is a behavioral scientist whose work focuses on the moments when customers are uncertain, skeptical, or under scrutiny, translating insights from consumer psychology into practical strategies that strengthen confidence, reduce friction, and improve adoption across products, services, and communication. His academic work is published in Financial Times Top 50 journals and has been covered by media outlets include Globe and Mail, Men’s Health, and the Wall Street Journal. He has conducted workshops and masterclasses for leading companies in Canada, Australia, and the United States, and has served on tenured faculties at Tyndale University, Toronto Metropolitan University, Purdue University, Monash University, and the University of Technology Sydney.
You’ll want to enrol if you want to…
-
… diagnose which governance failure—approval bias, escalation of commitment, the incentive gap, or diffused ownership—is occurring in a specific AI approval or oversight process, rather than treating every failure as “bad luck” or “bad actors.”
-
... distinguish an approval decision made on evidence from one made on confidence, and identify the deference or pilot-stage optimism driving the difference.
-
... set kill criteria before an AI initiative launches, and recognize when a good-faith review is actually a sunk-cost defense of a past decision.
-
... identify the cost/visibility asymmetry that keeps bad news about an AI initiative from reaching leadership, and redesign the reporting path to close it.
-
... assign real ownership to AI governance decisions, distinguishing accountable committees from ones where responsibility is diffused across so many people that it belongs to no one.
-
... apply one diagnostic sequence across all three courses in the package—person, team, and organization—to determine where in the stack an AI initiative is actually failing.