Corporate Business Alliance

CBA-AIG · Study material

Contents

The complete contents of the CBA Certified AI Governance Professional study material: every chapter and lesson, the reading time of each, and where the assessed workbooks fall. One lesson is open to read in full.

Lessons

99

Reading time

5 hours

Workbooks

20

Words

64,259

These figures are counted from the material itself. The reading time assumes a steady pace and no re-reading, and does not include the time spent building the workbooks.

Free, no account

Read a complete lesson

Why the disciplines you already run do not cover it, from Foundations of AI Governance, as a candidate reads it, slides included.

Read the lesson

Contents

7 chapters, 99 lessons. Lessons marked workbook carry a spreadsheet you build and submit, marked against the published tolerance. The award requires 6 of the 20 completed to that tolerance.

Introduction

1 lesson · 3 min

  1. CBA-AIG Study Guide3 min

01Foundations of AI Governance

14 lessons · 47 min

  1. What this chapter covers2 min
  2. The case organisation and the figures this guide is built from2 min
  3. What AI governance is for3 min
  4. Why the disciplines you already run do not cover itFree3 min
  5. Why a regime is needed at all: the specific things that go wrong5 min
  6. Defining an AI system for the purpose of an inventory5 min
  7. Where the boundary sits: four cases that decide most inventoriesworkbook4 min
  8. The AI lifecycle and where governance attaches4 min
  9. The three lines of defence applied to AI3 min
  10. Roles: who owns, who validates, who approves, who answersworkbook3 min
  11. Building and maintaining an inventory that stays trueworkbook6 min
  12. Frameworks: learn the structure, confirm the provisions3 min
  13. Common pitfalls2 min
  14. Chapter summary2 min

02Risk Classification and the Regulatory Landscape

22 lessons · 79 min

  1. What this chapter covers2 min
  2. Classification follows the use, not the technique4 min
  3. The shape of a tiered regime, taught structurally5 min
  4. The landscape Kelso actually sits in4 min
  5. The factors that push a use up a tier4 min
  6. Classifying Kelso's eight systems5 min
  7. The two contested classifications, and how to defend themworkbook6 min
  8. Kelso's second problem: everything is high risk, so nothing is7 min
  9. One technique, two uses, and the reclassification triggerworkbook5 min
  10. Four instruments, and why they are not four versions of the same thing2 min
  11. The EU AI Act: a product-safety regime wearing AI clothing4 min
  12. The NIST AI Risk Management Framework: a vocabulary you can borrow2 min
  13. ISO/IEC 42001: proof that the system exists, not that the model is good2 min
  14. The UK approach: existing regulators, existing powers2 min
  15. Extraterritorial reach: why a Leeds head office is not an answerworkbook2 min
  16. The rules that already apply and never mention AI3 min
  17. The obligations that recur whatever the label4 min
  18. Roles along the supply chain, and what buying does not buy youworkbook6 min
  19. Building a compliance map when the rules are still movingworkbook4 min
  20. What to do about the parts that are not settled2 min
  21. Common pitfalls2 min
  22. Chapter summary2 min

03Governance Frameworks and Management Systems

14 lessons · 52 min

  1. What this chapter covers2 min
  2. What a management system is, and why the shape keeps recurring4 min
  3. Designing a policy that can be testedworkbook6 min
  4. Standards, procedures and guidance, and why the distinction is load-bearing2 min
  5. The model lifecycle as a control frameworkworkbook8 min
  6. Gates are only real if somebody can fail one2 min
  7. Proportionate control, and why the difference between tiers has to be realworkbook5 min
  8. Third-party and procurement governanceworkbook4 min
  9. What you cannot outsource2 min
  10. Assurance: four different things doing four different jobs6 min
  11. Metrics for the governance function itself3 min
  12. Making it survive: review, triggers and decay3 min
  13. Common pitfalls2 min
  14. Chapter summary3 min

04Documentation, Transparency and Accountability

15 lessons · 44 min

  1. What this chapter covers3 min
  2. The sufficiency test: could a competent stranger reconstruct this4 min
  3. What a model record contains, and why each field is thereworkbook4 min
  4. Documenting data: provenance is a chain, not a label3 min
  5. Decision records: the reasoning outlives the decision3 min
  6. Transparency is plural: five audiences, five different artefacts3 min
  7. Meaningful information about the logic involved3 min
  8. Explaining a model against explaining a decisionworkbook3 min
  9. Notification and disclosure: when a person should be told3 min
  10. Contestability: a route to challenge that worksworkbook3 min
  11. Accountability: one name, and why a committee cannot be it3 min
  12. Documentation you did not write: bought-in systems and vendors3 min
  13. Keeping documentation alive: versions, retention and review2 min
  14. Common pitfalls2 min
  15. Chapter summary2 min

05Bias, Robustness and Testing Concepts

18 lessons · 42 min

  1. What this chapter covers2 min
  2. Fairness is a claim about values, not a property of a model2 min
  3. Kelso's fifth problem: unawareness is not fairness4 min
  4. Proxies: how a postcode carries what a protected characteristic would have2 min
  5. The main fairness measures, in ratios3 min
  6. Why you cannot satisfy them all: the impossibility when base rates differ2 min
  7. The label is not the truth2 min
  8. Choosing which measure governs is a values decision with a named approver2 min
  9. Testing for bias when you do not hold protected characteristic data3 min
  10. Small subgroups, statistical power and the pre-registered minimum cell sizeworkbook2 min
  11. Robustness: shift, adversarial input, edge cases and rare groups2 min
  12. Drift: a model validated once is unvalidated nowworkbook2 min
  13. Testing regimes: what before, what after, and how often4 min
  14. Independence: the builder cannot be the validator2 min
  15. Reading a validation report criticallyworkbook2 min
  16. What the frameworks expect of testing2 min
  17. Common pitfalls2 min
  18. Chapter summary2 min

06Incident Handling and Organisational Structures

15 lessons · 36 min

  1. What this chapter covers2 min
  2. What counts as an AI incident, and why the definition must be written first4 min
  3. Detection: how an AI failure actually surfacesworkbook2 min
  4. Triage and severity: scale, harm, reversibility2 min
  5. Containment: four options, each with a priceworkbook3 min
  6. Communication: audiences, clocks, and the sequencing problem3 min
  7. Redress: putting people back, including those who did not complain2 min
  8. Root cause analysis when the behaviour was learned rather than written3 min
  9. Post-incident review and the actions that actually prevent recurrence2 min
  10. Where the AI governance function sits, and what the three lines must actually do3 min
  11. The review board: what it is for and how it fails2 min
  12. The board, its committees, and the cadence problem2 min
  13. Culture, escalation, and the safety of the person who raises the concern2 min
  14. Common pitfalls2 min
  15. Chapter summary2 min

What the material is

The study material is written to the discipline’s domains and learning objectives, chapter by chapter, with worked examples and unscored checks in the lessons and assessed workbooks where the level requires applied work. It does not contain a bank of practice questions matching the examination; the specimen paper shows the style and standard of the examination, with a written rationale for every option.

Back to CBA-AIGCertification scheme