1. Purpose
Our organization uses approved artificial intelligence tools to support work while protecting people, confidential information, intellectual property and the quality of our decisions.
2. Approved use
Employees may use only approved tools and accounts for authorized tasks. AI output is a draft or input to human work, not an independent decision or verified source.
3. Prohibited information
Do not enter personal, confidential, privileged, regulated, security-sensitive or proprietary information unless the organization has expressly approved the tool and use case.
4. Human responsibility
The employee remains responsible for accuracy, fairness, source checking, tone, approvals and the final work product.
5. Escalation
Employees must pause and ask the designated policy owner when a use case involves sensitive data, consequential decisions, unfamiliar tools or uncertainty.
Test the policy before rollout
Ask employees and managers to apply the draft to several ordinary tasks. If two reasonable readers reach different answers, clarify the approved tool, information category, reviewer or escalation path. Record unresolved legal and privacy questions for qualified advisers rather than hiding them under general wording.