AI on campus: assistants that respect permissions and keep people in charge
AI assistants can answer campus questions quickly, but only safely if they respect who is asking. Principles for AI that is useful, grounded and accountable.
Institution leaders are being asked about AI from every direction: by boards, faculty, parents and vendors. The interest is understandable. A well-designed assistant can answer routine operational questions in plain language, saving staff from running reports or searching through files.
But an educational institution holds sensitive information about children, young adults, employees and finances. Adopting AI responsibly means asking not only what an assistant can do, but what it should be allowed to see, how its answers can be checked and who remains accountable for decisions.
Permissions must come first
An AI assistant should never become a shortcut around access controls. If a class teacher cannot open another section's marks in the ERP, they should not be able to obtain those marks by asking the assistant. If a parent can see only their own child's records, the same must hold in every conversation.
In practice, this means the assistant must work within the same role-based permissions as the rest of the system, establishing who is asking and what they are entitled to see before it answers. The same applies to actions: if the assistant can help start a request or prepare a report, it should do so only within what the user is already permitted to do.
An assistant that ignores permissions is not a productivity tool. It is a data leak with a friendly interface.
Ground answers in institutional sources
General-purpose AI models can produce confident answers that are simply wrong. On campus, an incorrect answer about an attendance policy or a fee deadline can cause real harm. Assistants should ground their answers in the institution's own policies and data, and show the sources they relied on, so that users can verify what they are told. Where no reliable source exists, it is better for the assistant to say so than to guess.
Keep people in charge of decisions
AI is well suited to finding, summarising and flagging information. It should not make decisions about students or staff on its own. A sensible set of principles for campus AI includes:
- Use AI to surface information and suggest actions, with a person confirming any change
- Treat predictive risk signals as prompts for human follow-up, not as verdicts
- Record AI-assisted actions in the same audit trail as other actions
- Be open with staff, students and parents about where AI is used
- Review how the assistant is used and adjust access as roles change
These principles are as much about governance as technology. Agreeing them before rollout, and revisiting them as use grows, helps build the trust that makes AI genuinely useful.
Where mycampace stands
The mycampace AI assistant is permission-aware: it answers natural-language questions within the user's role and context, grounds its responses in institutional policies and data, and shows its sources. Alongside predictive analytics, BI dashboards and natural-language queries, it is designed to help people find and act on information faster, while decisions stay with them.
