AI Governance

Introduction

AI governance defines who can deploy what model with which prompts, with audit, review, and compliance baked in.

Beginner analogy: Like change management for IT — approvals, audits, rollback.

Understanding the topic

Core concepts to understand:

• Roles: owners, reviewers, approvers.
• Risk tiers: low / medium / high — different review depth.
• Catalog: every prompt registered with metadata.
• Audit trail for regulators (EU AI Act, NIST AI RMF).

Syntax reference

Visual workflow / architecture:

Prompt change
  ▼ tier (low/med/high)
  ▼ review (1–3 people)
  ▼ eval pass
  ▼ deploy
  ▼ audit log

Real-world use

Banks & insurers mandate AI governance reviews for any LLM-powered feature.

Best practices

• Tier by risk; don't review every prompt the same way.
• Automate eval pass as approval gate.
• Keep an immutable audit log.

Hands-on exercise

Interview preparation — practice these questions:

• Q1. What is AI governance?
• Q2. Risk tiers — examples?
• Q3. Audit log fields?
• Q4. EU AI Act implications?