Before you paste anything into an AI tool, ask one question: “What is the most sensitive piece of information in what I’m about to enter?” Then match it to the table below. Your member classifies data using the System Cybersecurity Standards; some members use additional local categories. When in doubt, treat data as more sensitive, not less, and ask your IT or security office.
| DATA SENSITIVITY | EXAMPLES | WHERE IT MAY GO |
| Public | Already-published web content, public reports, course catalogs, press releases, public directory data. | Any tool, including consumer AI — but still verify outputs. |
| Internal / University-Internal | Routine internal email and memos, draft documents, non-sensitive operational data, internal procedures. | Approved enterprise tools only. Never consumer tools. |
| Confidential | Student records (FERPA), personnel files, donor and financial data, attorney–client material, security details, anything personally identifiable. | Only enterprise tools explicitly approved for that data level. Confirm with your IRM first. |
| Critical / Highly Regulated | Health information (HIPAA), Social Security and government ID numbers, payment-card data, export-controlled (ITAR/EAR) material, law-enforcement data. | Do not enter into general AI tools. Requires a specifically approved, contracted system and IRM/security sign-off. |
Quick test
If you would not post it on a public website, do not put it into a consumer AI tool. If it identifies a person, protect it like it identifies you.