Safe AI use checklist

Remove customer, employee, financial, legal, health, credential, and confidential company information unless the tool is approved for that use.

Ask whether the same result could be achieved with public, anonymized, or simplified context.

Check whether your organization has approved tools, policies, or client requirements that apply.

Treat AI output as a draft, not a source of truth.

Check facts, calculations, citations, assumptions, and missing context.

Look for confident wording that may hide uncertainty or unsupported claims.

Decide who owns the final answer, decision, or recommendation.

Use human approval for customer-facing, team-facing, or decision-facing work.

Keep enough notes, sources, or prompts that someone else can understand how the result was produced.