Module 11 Activity: AI Use Safety Review

Activity ID: AIT-M11-A01 Estimated active time: 35-50 minutes

Purpose

Apply privacy, bias, human-review, and recordkeeping checks to one safe workflow.

Use safe, fictional or public material throughout. Nothing in this activity requires a real name, account, record or figure.

Activity Steps

Part 1: Choose one nonsensitive workflow.

Use a real workflow, but one whose content is safe. The structure is what you are testing.

Part 2: List information that is safe, anonymised, or not shareable.

Three groups. If anything lands in 'not shareable', say what you would do instead of using the tool.

Part 3: Mark possible bias or unfair-assumption risks.

Look for anything guessing at motive, ability or character, and run the swap test on the wording.

Part 4: Define human-review triggers.

Name the category and the reviewer as an actual role. 'Someone should look at it' is not a trigger.

Part 5: Write a short AI-use record and stop conditions.

Record what AI helped with, what was checked, what was not, and who reviewed it.

What good looks like

A strong response includes at least one thing the review stopped, and says what was done instead.