Course and programme enquiriesforensics@siftcon.africa

AI risk and assurance · Technical and control teams

Turn AI principles intotestable control evidence.

Review AI systems for oversight, data, privacy, security, model and supplier risks using practical questions and evidence.

Duration3 days
LevelIntermediate
FormatCase-led lab
Learning units15

Learning outcomes

Skills you can use and explain.

Every result combines practical use, human review and evidence that another person can check.

  1. 01

    Scope an AI system, use case and accountable lifecycle owners.

  2. 02

    Map risks to people, organisation and society using a consistent taxonomy.

  3. 03

    Assess data, privacy, security, fairness and information-integrity controls.

  4. 04

    Challenge vendor and foundation-model dependencies.

  5. 05

    Design tests, evidence requirements and issue ratings for assurance.

  6. 06

    Produce an AI assurance plan and executive-ready opinion.

Course pathway

Govern. Map. Measure. Manage. Assure.

Five modules follow AI risk from the first system review to a clear final report.

Module 01System context and accountability
  • Assure a system, not a label
  • AI system context map
  • Scope decision
Module 02Risk and impact mapping
  • Connect harm to mechanism
  • AI risk taxonomy lab
  • Risk-statement decision
Module 03Data, security and model controls
  • Controls need evidence and thresholds
  • Control evidence test
  • Evidence sufficiency decision
Module 04Third-party and change risk
  • Know what depends on the vendor
  • Vendor assurance challenge
  • Material-change decision
Module 05Assurance planning and reporting
  • Conclude only within the evidence
  • AI assurance plan capstone
  • Assurance-opinion decision
SIFTCON Academy practical learning environment.

Final review project

Build an AI review plan based on evidence.

Learners define the scope, risks, control goals, tests, evidence, issue ratings and report for a realistic AI process.

ScopeTestChallengeConclude

Designed for

People who need to make better technical decisions.

Best suited to

  • Risk, compliance, privacy and internal-audit professionals
  • Cybersecurity, data-governance and technology-assurance teams
  • Model-risk and responsible-AI practitioners
  • Product and procurement teams accountable for AI controls

Preparation

  • Working knowledge of risk or control assessment
  • One existing or proposed AI use case
  • Access to relevant policy, architecture and vendor materials
  • No model-development experience is required
01

System scope

Map components, people and accountable owners.

02

Risk register

Connect harms, causes and controls.

03

Control test

Examine design and operating evidence.

04

Assurance opinion

Report confidence, gaps and action.

Group and in-house training

Adapt this course to your learners, organisation and decisions.

Plan a cohort
Talk to the AcademyWhatsApp us
WhatsApp