Data Privacy and Security

The Core Commitment

AI tools are widely available and many people across Maricopa are eager to use them. This page covers how to do that without putting institutional, personal, or proprietary information at risk — and who to involve when a situation is bigger than one person’s judgment.

Data privacy and security is one of the five GUIDE principles, and the one that makes the others possible: verified security is what allows people to experiment without fear. Clear protection of personal and institutional data is not a brake on innovation, it is the condition for it.

Information or data that never goes into an AI tool

Five categories, regardless of the tool:

  • Student information — protected under FERPA. Grades, contact information, academic records, and anything that could identify a specific student.
  • Personnel information — employee records, performance evaluations, salary data, disciplinary information, and other HR content.
  • Confidential institutional data — anything Maricopa has promised to keep confidential, such as contract details, grant applications, or research participant disclosures.
  • Non-public institutional information — anything confidential, proprietary, or not yet intended for public release.
  • Intellectual property — yours, the district’s, a co-author’s, or a third party’s. Submitting it may surrender rights to control how it is used.

Responsible and Safe Use Basics covers information security in more everyday terms, and Privacy Considerations covers how specific tools handle data.

Seek Assistance When Needed

Some AI work triggers requirements that already exist in district policy. These are not new ARC rules — they are the same gates that apply to any system, and AI does not exempt anything from them. The point of listing them here is so you meet them at the start rather than discovering them at the end.

Before anything else, establish what kind of data is involved. The five categories above are the short version; the district’s data classification standard is the authority. If the data is protected, that determines which tools are permissible, what agreements must be in place, and whether the activity can proceed at all.

When you are not sure how something classifies, ask before you proceed rather than after.

A tool the district has not reviewed has no agreement covering what happens to what you put in. This applies to free tools as much as paid ones — a free product processing institutional data is not exempt because no money changed hands.

If you want to use something that is not on the approved list, that is a conversation with IT before you start using it, not after.

Acquiring an AI tool, or adopting one at department or college scale, goes through the district’s normal contract and procurement processes. Signature authority rules apply. An individual agreeing to a vendor’s terms of service on behalf of a department is a contract, whether or not it felt like one.

If what you are building will be used by students or the public, accessibility requirements apply to it. AI-generated output frequently looks accessible without being accessible — see Accessibility for what to check and why automated output needs human review.

Arizona records retention requirements apply to institutional records regardless of the system they live in. If an AI tool is being used as part of an official process — advising, personnel, student services — consider what is being created, whether it constitutes a record, and how it is retained.

This is the requirement people most often miss, because a chat window does not feel like a filing system.

Sources and Document History

If you are unsure which of the above concerns applies to something you are planning, Choosing the Right Tool works through it in a few minutes, or contact the ARC at ai-resource-center@domail.maricopa.edu and you will be pointed to the right office.

District regulations and guidance