AI Guidelines for Staff
Evaluation and Risks
Institutional safety and efficiency depend on the responsible use of AI in our administrative and operational workflows. Below are the key requirements for all staff members.
The "Must-Know" Essentials
- Mandatory Vetting: All AI tools, including those used for research or integrated into existing software, must be vetted and approved by IT before use.
- 30-Day Lead Time: Submit your request to IT at least 30 days before you plan to use a tool to allow for security and compliance reviews.
- No Automated Decisions: AI must not be used for automated decisions (like grading or admissions) without direct human oversight.
- Protect Student Privacy: Never input sensitive or personally identifiable information (PII), such as student grades or records, into AI tools unless IT has specifically approved that tool for such data.
- Institutional Accounts: Faculty and staff may use non‑institutionally provided AI tools for work‑related purposes as long as no institutionally protected data is entered. Protected data includes personally identifiable information (PII), student records covered by FERPA, employee data, and other confidential College information. Tasks involving protected data must use institutionally approved AI tools.
Operational Responsibilities
- Verify All Outputs: AI can produce "hallucinations" or fabricated facts; always use your professional judgment to verify AI-generated content.
- Disclosure: Disclose the use of AI in administrative outputs as required by institutional policy.
- Contractual Awareness: AI-integrated tools are subject to risk assessments upon contract renewal.
- Monitor for Changes: Tools will be re-evaluated if the vendor updates terms of service or releases major new AI features.
Safety and Compliance
- Immediate Reporting: If you accidentally upload restricted data or notice a tool requesting information beyond its purpose, report it to the IT Help Desk immediately.
- "No-Fault" Culture: Our priority is data protection and risk containment; reporting an accident will not result in fault.
- Compliance Risks: Non-compliance may lead to loss of tool access, reporting to supervisors, or disciplinary action.
What to Do If You Make a Mistake?
We maintain a "no-fault" reporting culture for accidental data exposure. If you accidentally upload restricted data (like student records) or notice an AI tool behaving unexpectedly, report it to the IT Help Desk immediately.
Why report? Your quick action allows the IT team to contain the risk and protect the community’s data.