AI Governance & Risk

Data and Model Boundaries

2 lessons · ~6 min

Estimated reading time.

Overview

Every governance program starts with two boundaries: what data is sensitive enough to need a rule, and which models are actually allowed to touch it. This module sets both before anything else in this track makes sense.

1
Classifying Your Data
~3 min
Names the specific tiers of data sensitivity a governance program actually needs — not a vague 'be careful with data' rule, but a classification anyone in the org can apply consistently before AI ever touches it.
2
The Model Allowlist
~3 min
Replaces 'whichever AI tool someone happens to use' with an explicit list of approved models and providers — the second boundary a governance program needs, right after data classification.