Writing an AI policy
An acceptable-use policy your team will actually follow: scope, approved tools, data rules, disclosure, and how to keep it current.
6 lessons · about 1 hour
What’s inside
- 01Why you need an AI policy
Staff are already using AI; a clear policy with a defined scope and usable principles turns quiet risk into guided practice. · 10 min
- 02Approved tools, and how to approve new ones
Naming which tools and plans people may use, and for what, turns judgement calls into a lookup and steers people to versions that protect your data. · 11 min
- 03Data rules: what never goes in
Clear, classified rules about which information can go into which tools, practical alternatives, and what to do when something goes wrong. · 10 min
- 04Disclosure, accountability and human review
When to tell people AI was involved, why a named person is always accountable, and how to make human review real rather than a rubber stamp. · 11 min
- 05Training, support and keeping it current
How to help people follow the policy, what the amended EU AI literacy duty asks, and a review cadence that stops the policy going stale. · 11 min
- 06A worked one-page acceptable-use policy
A complete worked outline of a one-page AI acceptable-use policy for a fictional charity, with the reasoning behind each choice and a rollout plan. · 12 min
