A strong AI policy for a college office fits on one page and answers seven questions: who it covers, which tools are approved, what information can go in, what good use looks like, what’s off-limits, how AI-assisted work gets reviewed and disclosed, and who to ask. Write it in plain language, so your staff actually read it and use it.
Most teams are already experimenting with AI, and there’s real time to gain in registrar, program planning and advising work. A clear guideline gives everyone the same starting line: what’s allowed, where to begin and how to protect the students you serve.
Start with the why now
Before you write a single rule, name the reason. People adopt a change when they understand the what, the why and the why now. Open your guideline with two or three sentences on why your office is putting it in place today, tied to the work your team does every week.
The seven parts of a one-page AI guideline
- Who it covers. Name each group: full-time staff, student workers, part-time staff, faculty who support the office. Naming them up front saves a lot of “does this mean me?” emails.
- Approved tools. List the AI tools your IT team has approved, one per line. Staff want a clear yes, and this tells them exactly where they can start today.
- A data traffic light. Sort the information your office handles into green (fine to use), yellow (only in an approved tool) and red (never goes in). People remember three colors far longer than a paragraph of rules.
- Good uses. Give everyday examples from your own office, like drafting a first version of a routine email, summarizing a long policy document or outlining a training agenda. People learn fastest from examples.
- Off-limits uses. Keep this list short so everyone remembers it. A few firm lines protect the people you serve and your staff.
- Review and disclosure. Say who checks AI-assisted work before it goes out, and when a message should note that AI helped. Trust grows when people know how a message was made.
- Who to ask, and when you’ll update it. Name a contact for questions and new tool requests, and set a review date: every term, every six months or every year. AI tools change fast, and a set date keeps your guideline current.
How should the guideline handle student data and FERPA?
Put student records in red by default. They stay out of AI tools unless your IT and compliance teams have approved a specific tool for that use, under your institution’s own agreements and security settings.
Then make green easy to use. Public information, your own drafts and templates, and anonymized group counts are where most offices can start today, with no student data involved.
Who should be in the room?
Invite the people who do the work. When the people affected by a change help shape it, the guideline fits how they actually work, and they’re far more likely to follow it. Then share your draft with your legal, compliance and IT teams for review before you adopt it.
A guideline sticks when people practice it
Writing the page is step one. Adoption comes from practice: your team working through real tasks inside these guardrails, with tools set up to match. Plan a short session where everyone tries a good use from the list on their own work, and check in after the first month to see what’s working.
One page, plain language, a clear yes. That’s a guideline your team will actually use.
Try it: My free AI Use Guidelines Builder walks you through all seven parts and builds your one-page guideline as you go, in about 15 minutes. Download it as a Word file or PDF and share it with your team. It’s a starting template, so have your legal, compliance and IT teams review it before you adopt it.
If you’d like your team to practice with the guideline on real work, that’s what my AI Enablement Workshop is for. Let’s talk.
