Executive Learning · Free learning
Ethics & Responsible AI
Fairness, transparency, accountability and the regulatory floor.
In this module
0 of 5 activities complete
Reading
0% readWhere harm enters
Harm enters through biased training data, proxy variables that stand in for protected characteristics, opaque decisions that cannot be explained to those affected, and automation applied to decisions that require human discretion.
The governance minimum
A named accountable owner, a documented purpose, a record of training data provenance, an explainability standard proportionate to impact, a human appeal route, and periodic drift monitoring.
The regulatory direction
Risk-tiered regimes such as the EU AI Act place the heaviest obligations on decisions affecting employment, credit, education and essential services. Assume disclosure and documentation obligations will only increase.
Ethics & Responsible AI, executive briefing
18:00 · Video briefing
Key takeaways
- Bias usually enters through data and proxy variables, not intent
- Explainability should be proportionate to the impact of the decision
- Every consequential automated decision needs a human appeal route
Resources and downloads
One-page AI use policy
Governance floor covering ownership, provenance, explainability and appeal.
Draft a one-page AI use policy
Set your governance floor before you scale.
- List decisions your organisation will never fully automate
- Define who signs off a model that affects a person directly
- Define how an affected person appeals a decision
Module 8 quiz
2 questions · 70% to pass
01A hiring model excludes candidates by postcode. This is an example of:
02Which control most directly protects an individual affected by an automated decision?Applied
Your notes
Lesson discussion
Ask questions, share how you are applying this in your organisation, and learn from other executives.
Sign in to join the discussion. Reading is open to everyone.
Loading discussion…