Your AI council will shape how thousands of people teach, learn, and handle student data. Fill these four seats, and IT stops carrying every decision by itself.
Most university AI councils spend their first year approving software requests. Meanwhile, the decisions that shape teaching, student data, and academic integrity sit unmade.
The pattern starts on day one. When a campus forms an AI council, the work usually lands in the CIO's office. AI runs on software, so the logic seemed obvious at the time. Then the real questions showed up, about grading, student records, vendor contracts, and cheating, and most of them had nothing to do with servers.
Here's how you can tell your council was built for the wrong job:
Faculty learned about the new AI guidance from the student newspaper.
Three colleges bought three overlapping AI tools, and nobody checked what happened to the data.
Questions about cheating, grading, and disclosure keep arriving as IT tickets.
The draft policy went to legal in the spring and hasn't come back.
Every seat at the table represents a box on the org chart, but nobody represents a classroom.
Each of these traces back to the same cause. The decisions belong to people who aren't in the room, so they drift to the one office that is... Yours. And no firewall rule will settle a grading policy.
The fix starts with who sits at the table. Four seats matter more than the rest, and the order matters too.
The Provost or Chief Academic Officer
Most AI questions on campus sound technical. Very few of them are.
Should students be allowed to use AI on a take-home exam? Does a faculty member have to disclose that a chatbot helped draft the syllabus? What counts as original work in a writing course now? Every one of those questions is about teaching and learning. Your team can host the tool, secure it, and connect it to the LMS. Nobody in IT can decide what it means for a grade.
The provost can.
What the Provost Decides That IT Can't
The provost's office owns the academic standards that AI touches every day. That includes course policies, assessment expectations, academic integrity rules, and what faculty need to know before they teach with AI. When those decisions sat with IT on other campuses, they stalled. Faculty heard guidance from the technology office and treated it as optional, because in their world, it was.
A provost at the table changes the weight of every decision. A recommendation about AI in assessment now carries the authority of academic policy. Deans pay attention, department chairs plan for it, and faculty see that the people who evaluate their teaching care about this too.
Share the Chair With the Provost
Here's where many councils go wrong. They invite the provost to the kickoff, get a supportive quote for the announcement, and never see that office again.
Ask for a shared chair instead. Split the agenda in two. You own infrastructure, security, procurement, and integration. The provost owns teaching, learning, research expectations, and faculty readiness. Each meeting should end with a decision from both sides.
If the provost can't attend every session, ask for a vice provost with real decision rights. A delegate who has to "take it back to the office" turns every meeting into a waiting room.
The Director of Teaching and Learning
Add the person who knows how faculty work.
The provost sets the standard. Someone still has to help a chemistry professor figure out what that standard looks like on a Tuesday morning in front of 140 students.
That person runs your center for teaching and learning.
Most councils skip this seat. The director rarely ranks high enough to make the org chart cut, and the work sounds like something to hand off after the policy is finished. That's backward. Policy written without this person tends to fall apart the first time it meets a real syllabus.
Why Guidance Without Practice Goes Nowhere
Think about how most campuses rolled out AI guidance over the last two years. A committee wrote a memo that went out by email. Faculty skimmed it between grading and committee work, and then they did one of three things. Some banned AI outright, some ignored it, and some quietly experimented on their own with no support.
None of those reactions came from bad faith. Faculty were handed rules with no practice attached. They knew what they weren't supposed to do and had no idea what good use looked like in their own discipline.
The director of teaching and learning solves that. This is the person who turns a council decision into a workshop, a sample syllabus statement, a redesigned assignment, or a faculty learning community that meets every other week. Without that translation, your policy lives on a web page nobody visits.
Ask the director to own three things for the council:
Faculty readiness. A clear picture of what faculty can do with AI today, by college or department, so the council plans from what's real on campus.
Practical resources. Syllabus language, assignment examples, and review checklists that faculty can adapt in an afternoon.
Feedback from the classroom. A standing report on what's working, what's confusing, and where faculty are stuck.
That last item carries the most weight. Your council will make better decisions when it hears regularly from the people trying to apply them.
A Quick Test for Your Current Council
Pull up your last three meeting agendas. Count how many items came from someone who taught a class this semester. If the answer is zero, this is the seat to fill next.
General Counsel or the Chief Privacy Officer
Remember the draft policy that went to legal in the spring? Odds are it stalled for a simple reason. Legal saw it last.
By the time a finished policy reaches the general counsel's desk, every risk question shows up at once. Who owns the data a student types into a chatbot? Does this vendor train its model on campus content? Can a faculty member paste a graded essay into a free tool? Each question sends the draft back for another round, and months disappear.
The Questions Only Legal Can Answer
Your team knows how data moves. Counsel knows what the institution is allowed to do with it. AI tools sit right where those two meet.
A few examples come up on almost every campus:
Student records. FERPA protects education records, and a lot of everyday AI use touches them. Grades, written feedback, advising notes, and student work can all end up in a prompt.
Vendor terms. Free and low-cost AI tools often come with terms that let the vendor keep or reuse what users enter. Departments that bought tools on a purchasing card rarely read those terms.
Accessibility. AI tools used in courses still have to work for students with disabilities. That obligation doesn't pause for new technology.
Research data. Grant agreements and IRB protocols may limit where research data can go, including into an AI model.
None of these questions has a technical answer. Your security team can block a tool. Only counsel can tell you whether the institution was allowed to use it in the first place.
How to Bring Legal In Without Slowing Everything Down
Invite counsel or the chief privacy officer as a standing member, starting with the first meeting. Then give them a specific job. Ask them to build a short list of data rules the council can apply to every new tool request, including which data is approved for which tools, which contract terms are deal breakers, and when a request needs a full legal review and when it can move ahead.
That list will do more for your team than any single policy. Procurement gets faster because the answer to most requests is already written down. Departments know the rules before they buy. Legal spends its time on the hard cases instead of every case.
The Payoff for IT
When legal sits on the council, your office stops serving as the final checkpoint for contracts nobody read. Your team enforces rules the institution agreed on together, and that's a much easier position to defend.
The Dean of Students
Count the students represented on your AI council. On most campuses, the number is zero.
That's a strange outcome for a group making decisions about tools students use more than anyone else on campus. Students are the ones submitting the work, sitting through integrity hearings, and trying to figure out why AI was fine in one course and banned in the next.
The dean of students hears those stories every week. That's why this seat belongs on your council.
What the Dean of Students Sees That Nobody Else Does
Academic integrity cases tell you more about your AI guidance than any survey. When cases climb in one college, the rules there are probably unclear. When a student says, "My other professor said this was fine," the council has a consistency problem to fix. The dean's office spots those patterns early, often a full semester before the numbers reach a report.
The dean also sees who gets left behind. Some students pay for premium AI subscriptions. Others work with whatever's free, sometimes on a phone between shifts. When a course assumes access to a paid tool, the playing field tilts, and the students on the low side tend to be the ones with the least room for error. A council that never hears this will make decisions that look fair on paper and land unevenly in real classrooms.
Bringing Student Voices In
The dean can also open a direct channel to students. Ask the dean's office to recruit one or two student representatives, possibly through student government, and to gather input before major decisions. Students will tell you things faculty and staff won't, including which tools they use and where the rules confuse them.
Give the dean a clear assignment on the council:
Integrity patterns. A short report each term on AI-related cases, grouped by college, with the confusion points that keep repeating.
Access concerns. A list of places where cost, devices, or connectivity put some students at a disadvantage.
Student input. A summary of what students asked for, worried about, or misunderstood since the last meeting.
A council is only as useful as the decisions its members can own. Seat the provost, the director of teaching and learning, legal counsel, and the dean of students, and the questions that kept landing in your queue will finally reach the people equipped to answer them.
Even the right council has a limit, though. A council can decide what's allowed. It can't teach a history professor how to write an AI brief for next week's lesson, or show an advising team how to use AI without exposing student records. Policy tells people where the lines are. Training shows them what good work looks like inside those lines, and most campuses have far more of the first than the second.
Our training and workshop options can be shaped around the decisions your council has already made, so faculty leave with something they can use in their next class, staff learn to apply AI to the work already on their desks, and leaders across campus share a common way of talking about AI.
Your council sets the direction. We help the people on your campus follow it with confidence.
If your council is just taking shape, or if it's been meeting for months and still spends most of its time on software requests, schedule a call with us. We'll talk through what your faculty and staff need, which training and workshop options fit your campus, and how to turn your council's decisions into everyday practice.