Purpose and scope
Who the policy applies to and what AI tools it covers.
AI policy and shadow AI
Safe AI Rules for the Workplace
A clear AI policy helps people understand what is allowed, what is prohibited, what requires approval and when human review is required.
Adapt this sample before workplace use
The downloadable Shadow AI policy is an educational sample, not a ready-to-adopt disciplinary procedure. Review it against employment contracts, existing policies, applicable awards or enterprise agreements, consultation duties, your jurisdiction and the circumstances of the workplace.
Warnings, dismissal and summary dismissal are serious, fact-dependent decisions. Follow a fair process and obtain independent Australian workplace-relations or legal advice before disciplinary action. The Fair Work Ombudsman guidance on performance and warnings and its dismissal guidance explain current minimum-workplace considerations. Sources checked 27 October 2026.
What is shadow AI?
Shadow AI is unapproved AI use at work.
It often happens because AI tools are easy to access, useful and fast. However, unapproved AI use can expose confidentiality, privacy, cyber security, intellectual property, accuracy and compliance risks.
Policy sections
What an AI policy should include.
Approved tools
Which AI systems may be used for work and for what purposes.
Prohibited uses
When AI must not be used, including unlawful or high-risk purposes.
Data rules
What information must never be entered into AI tools unless controls are approved.
Human review
When outputs must be checked before publication, decisions or customer communication.
Accuracy and bias checks
How staff verify output and consider fairness.
Record keeping
How AI use is documented for important decisions or high-risk processes.
Incident reporting
How staff report errors, privacy concerns, security issues or harmful outputs.
Training and review
Minimum awareness training and policy review cycles.
Simple employee rule
If a tool is not approved, do not use it with work information. If the output affects a person, decision, customer, safety issue or legal obligation, a human must review it.