# Privacy, safety, and AI governance basics

> Create reasonable safeguards before scaling AI use.

Canonical: https://ainextia.com/learn-ai/privacy-and-governance/
Publisher: https://ainextia.com/

Level: Intermediate
Track: Evaluation
Reviewed: 2026-11-05

## Part 1
Classify inputs by sensitivity before choosing a tool. Check vendor retention, access controls, data residency, and whether prompts may be used to improve models. Make rules clear to staff.

## Part 2
Set approval gates for consequential decisions, data changes, external communication, and payments. Log exceptions and review incidents instead of assuming automation will always work.

## Part 3
Write a one-page AI usage policy listing allowed data, disallowed data, approved tools, and escalation paths.

## Knowledge check
What is the first control you would apply to an agent with access to customer records?

Reference: https://www.nist.gov/itl/ai-risk-management-framework
