Committee hears AI governance briefing; staff point to an imminent AI hub, training and guardrails
Loading...
Summary
University of Minnesota board staff briefed regents on emerging AI governance practices, said the university is establishing a university-wide AI hub and vice provost for AI, and recommended guiding principles, tool approval and training to manage data privacy and IP risks.
Board office staff told the Governance and Policy Committee that artificial intelligence is already affecting teaching, research and university operations, and that the institution is preparing governance and training resources to manage opportunities and risks.
Committee and policy coordinator Maggie Marchesani summarized peer approaches across the Big Ten and beyond, noting that some universities have adopted guiding principles, approved specific licensed AI tools for institutional data, and created multilayered governance councils. Marchesani said the university has named a vice provost for artificial intelligence and will launch a university-wide AI hub to centralize education, operational guidance and research. "From what we have found, you can take comfort in knowing that the university is at the forefront of AI," she told the committee.
Staff identified governance questions for the board: who will oversee AI at the institution, what guardrails are necessary to protect data privacy and intellectual property, how to manage change across units, and which board functions could benefit from AI (examples included searching minutes, synthesizing long dockets and preparing meeting memos). The board office suggested low-risk pilots for administrative use of licensed AI tools while retaining human review of outputs.
Regents raised practical questions about baseline training before tool access and confidentiality risks. One regent asked whether university-licensed AI tools require mandatory training and suggested controls that block access until training is completed. Staff replied that the AI hub will include upskilling and training for faculty, staff and the board office, and that human review and carefully scoped licensed tools are central to reducing legal and privacy risks.
Speakers emphasized the risks of placing board or attorney-provided confidential information into open-source models: "Anything they put into an open source AI is now the public domain," the committee chair said, warning that case law is still developing about privilege and data protections. Staff said they will continue benchmarking with peer institutions and professional networks (including the Association of Governing Boards) as best practices emerge.
The committee closed the discussion after questions and thanked staff for the briefing; there were no formal board decisions taken on AI governance during the meeting.
