Protect
Strip private, regulated, and confidential details first.
Responsible AI training
Question the answer. Check the evidence. Own the decision.
Practise the decisions behind responsible use: whether AI belongs in the task, which facts can be shared, what needs verification and when to hold or escalate the work.
Safe AI decision habit
Policy becomes practice.
Strip private, regulated, and confidential details first.
Give context, limits, and the decision standard.
Check facts, assumptions, bias, and missing context.
Move high-impact or uncertain work to a person.
Before someone uses AI
Risk and response map
Do not paste private, regulated, customer, employee, or confidential information into an AI tool without authorization.
Treat fluent output as a draft or suggestion until it is checked against context, facts, and policy.
Know when AI can assist and when a person must decide, approve, or escalate.
Do not copy AI output into real work before checking source quality, assumptions, and missing context.
Governance habits
01
Name the data boundary before prompting.
02
Ask for uncertainty, assumptions, and review points.
03
Keep human approval visible on decisions that matter.
04
Document what was checked before output was used.