Secure AI Use at Work Training

Approved Tools, Accounts and Use Cases

Identify who approves an AI service and the intended business task. Check which account, plan and policy apply, instead of assuming a consumer account has the same arrangements as an organisational service.

Break

Deciding What Information Can Be Entered

Practise removing personal identifiers, client details and business secrets from a fictional request. Check whether a task can be completed with a synthetic example or a smaller amount of approved information.

Checking Answers Before Using Them

Review a plausible but unsupported output. Verify important claims against authoritative sources, preserve human approval and consider the consequences of using an incorrect summary, calculation or recommendation.

Lunch

Files, Connectors and Instructions in Content

Discuss why an assistant reading a document or connected service needs appropriate permissions. Recognise that content can contain misleading instructions, and avoid granting extra access simply to complete a task quickly.

Break

Sharing Outputs and Reporting Problems

Check whether generated material contains confidential information or unverified claims before sharing. Identify the owner of retention and access settings, and report accidental disclosure or unexpected actions through the approved route.

Practical Exercise: Rewrite a Risky AI Request

Turn a fictional customer-support prompt into an approved, minimal-data request. Review the resulting draft for unsupported statements, identify the human approver and create a team safe-use checklist.