AI curriculum

Course outline for AI Governance

A policy has little operational value when nobody owns its exceptions or verifies its use. Create a fictional AI pilot dossier that links risk, delegated authority, human oversight, and evidence. Walk a policy change through distinct proposal, approval, deployment, observation, and revision responsibilities.

Technology: AI Security, Governance and Trust
Duration: 6 hours
Modules: 3 · 2 hours each
Level: Level 2

About AI Governance

AI Governance Course Objectives

Classify a fictional use case using explicit risk criteria.
Translate a policy into an authority matrix and observable checkpoint.
Assign owners for approvals, exceptions, review queues, and policy changes.
Specify transparency, evidence, and retention needs for a pilot.
Reject a policy-change proposal with missing accountability or unsupported evidence.

Pre-requisites

AI capability and limitation awareness equivalent to Artificial Intelligence Foundations.
Familiarity with organizational decision processes and accountable ownership.
Coding is not required; responsible-AI and human-workflow study are useful preparation, not enrolment gates.

Before this course

Complete the required courses or bring equivalent experience. Recommended courses add useful depth; they are not enrolment gates.

Required preparation

Artificial Intelligence Foundations

Recommended preparation

Understands AI use cases and organizational decision ownership; coding and full technical-security training are not required.

Find your place in the learning map

Lab Setup

Computer with a document editor or spreadsheet.
Use fictional organizational roles, synthetic pilot records, supplied governance-framework summaries, and guided policy walkthroughs.
No code, model account, compliance subscription, or real organizational data is required.

Detailed Course Outline

Module 1: Risk and operational policy

Governance frameworks: Compare supplied governance lenses for risk, ownership, and oversight; translate relevant concepts into pilot decisions without claiming framework conformity.
Risk classification: Classify a fictional use case by consequences, affected parties, and action authority; record criteria and uncertainty instead of assigning unexplained labels.
Policies: Translate organizational intent into allowed actions, evidence requirements, and exception conditions; separate operational rules from unsupported conformity claims.

Practical outcome: Compare supplied governance lenses and produce a risk record whose policy states allowed actions, required evidence, and exception conditions rather than claiming framework conformity.

Module 2: Oversight and controlled change

Human oversight: Specify reviewer authority, intervention points, queue capacity, and escalation; identify when available oversight cannot keep pace with incoming work.
Accountability: Assign owners for consequential actions, exceptions, and unresolved risks; distinguish responsibility for decisions from responsibility for operating each control.
Policy-change lifecycle: Separate proposal, approval, deployment, observation, and revision responsibilities; reject incomplete change requests and retain the evidence behind each transition.

Practical outcome: Assign separate proposal, approval, deployment, observation, and revision owners; model a review queue and escalation when oversight capacity is exceeded.

Module 3: Reviewable pilot evidence

Transparency: Identify what affected stakeholders need to know about the pilot, its limits, and decision responsibilities; distinguish useful disclosure from unsupported reassurance.
Auditability: Specify reviewable action and approval records, evidence ownership, and retention needs; make consequential decisions reconstructable without assuming legal compliance.
Evidence: Assemble a fictional pilot dossier linking risks, authority rules, and control gaps; distinguish documented observations from assumptions and unresolved requirements.

Practical outcome: Assemble a noncoding pilot dossier with stakeholder communication, approval records, evidence owners, retention needs, and unresolved control gaps.

Practical exercise

Submit a fictional pilot dossier, authority matrix, and policy-change walkthrough.
Compare a complete change request with missing-owner, missing-approval, and overloaded-review cases.
Give every consequential action an accountable owner, an approval or permitted-action rule, and a reviewable evidence requirement.
Reject all incomplete change fixtures, route the capacity exception to a named owner, and record unresolved risk before recommending whether the pilot may proceed.

What you will produce

Create a reviewable AI pilot governance dossier with accountable policy-change owners.

How the work is checked

Every consequential action has an owner, authority rule, and evidence requirement.
All incomplete change fixtures are rejected and review-capacity exceptions are escalated.
The dossier distinguishes conceptual governance from framework or legal conformity.

Six-hour noncoding pilot design; framework adoption, organization-wide rollout, and jurisdiction-specific obligations require separate assessment.

When another approach is better

Do not substitute this pilot dossier for an organization's binding policy or specialist legal advice.
A review checkpoint is unsuitable when no qualified reviewer has capacity or authority to intervene.
A low-risk deterministic workflow may need existing governance rather than a new AI-specific process.

How we train

Ask about exercises, instructor feedback and the prior knowledge you need. Tool-specific courses marked provisional may become modules in a broader course.

Discuss your learning goals