Skip to content
AfriScienceGroup home
AI and Digital Transformation

Responsible AI: Bias, Fairness and Assurance

A model that performs well on average can still be systematically wrong about a particular group of people, and the organisation deploying it carries the consequence whether or not it understood the risk. This course covers finding and managing that. Delegates work through where bias enters, from the problem definition through the training data to the deployment context, and the fairness measures available with the trade offs between them. Testing a model for disparate outcomes is practised. Explainability is covered honestly, including the limits of current techniques. Human oversight and the conditions under which it is meaningful rather than nominal are addressed. The course closes on documentation, assurance and the emerging regulatory expectations.

3 day courseRegister NowRequest Proposal

Course objectives

  • Identify where bias enters, from problem definition to deployment
  • Apply fairness measures and understand the trade offs between them
  • Test a model for disparate outcomes across groups
  • Use explainability techniques and state their limits
  • Design human oversight that is meaningful rather than nominal
  • Document a model so it can be assured by someone else
  • Anticipate the emerging regulatory expectations

Who should attend

  • Data scientists and machine learning practitioners
  • Risk, compliance and audit staff assessing models
  • Product owners deploying automated decisions
  • Legal and policy staff
  • Executives accountable for automated decisions

Course outline

  1. 01Right on average, wrong about a group
  2. 02Where bias enters
  3. 03Training data and historical outcomes
  4. 04Fairness measures and their trade offs
  5. 05Testing for disparate outcomes
  6. 06Explainability and its limits
  7. 07Meaningful human oversight
  8. 08Documentation, assurance and regulation

Scheduled sessions

Scheduled sessions for Responsible AI: Bias, Fairness and Assurance
DatesVenueFormatPriceRegister
21 to 23 October 2026OnlineOnlineR6,500 per delegateRegister Now
14 to 16 December 2026Dar es Salaam, TanzaniaClassroomUSD 995 per delegateRegister Now
22 to 24 March 2027Cape Town, South AfricaClassroomR14,950 per delegateRegister Now

Responsible AI: Bias, Fairness and Assurance

From R6,500

AskRegister