• Drafting and editing internal communications, general correspondence, and marketing content using non-confidential inputs.
• Summarizing publicly available or internal-tier documents to save review time.
• Research assistance — using AI as a starting point for research, with output independently verified before being relied upon.
• Code assistance for non-sensitive internal tooling, subject to normal code review.

• Submitting regulated or confidential data — client records, protected health information, financial account numbers, credentials, or trade secrets — to any AI tool not specifically approved for that data tier.
• Using AI-generated output for high-stakes decisions (hiring, termination, credit, medical, legal, or safety-related determinations) without meaningful human review.
• Presenting AI-generated output as final work product without verifying its accuracy.
• Using unapproved or personal-account AI tools to process organizational data of any classification above public.

• Only submit data to an AI tool consistent with the data classification rules in the AI Governance Policy.
• Do not paste customer, patient, or employee personal data into a general-purpose AI chat interface unless it is contractually confirmed not to train on submitted data and is on the approved list for that data tier.
• Remove or redact identifying details before using AI tools for illustrative or exploratory purposes wherever possible.

• Disclose AI assistance when it materially shaped a work product delivered to a client, regulator, or external party, consistent with any contractual or professional obligations.
• Disclose AI use internally when AI-generated content forms the basis of a decision, recommendation, or analysis presented to leadership.
• When uncertain whether disclosure is required, disclose — silence is the higher-risk default.

• Violations are handled under existing organizational conduct and data protection policies, up to and including termination for serious or repeated violations.
• Accidental disclosure of confidential data to an AI tool must be reported immediately so it can be assessed and, where required, escalated as a data incident.

By using any AI tool on behalf of the organization, I acknowledge that I have read and understand this Acceptable Use Policy, and agree to comply with its permitted uses, prohibited uses, and data handling and disclosure requirements.

• Employee name, signature/acknowledgment, and date.
• Policy version acknowledged, for version-tracking purposes.

• AI Acceptable Use Policies — /knowledge-center/artificial-intelligence/ai-governance-secure-ai/ai-acceptable-use-policies
• Shadow AI Explained — /knowledge-center/artificial-intelligence/ai-governance-secure-ai/shadow-ai-explained