The Provost’s Case for Institution-Wide AI

AI in higher education has already arrived on your campus. Whether it improves student outcomes depends on something most institutions haven’t tackled yet: getting faculty aligned and connecting what they do to what students actually learn.

Two-thirds of institutions have adopted AI.¹ For an academic leader, that number is deceptively reassuring. It says the technology has arrived. But it says nothing about whether it’s making students better writers, thinkers, and learners.

That distance between AI being present in higher education and AI benefitting students is where the real work of an academic leader begins. Access isn’t adoption. Even where people are genuinely using AI, that use only moves student success when two things are in place—faculty who are aligned and outcomes you can actually see.

The alignment problem

Walk across most campuses and you’ll find the same pattern: AI use that varies wildly from one classroom to the next, and not because faculty disagree about its value.

The reason is usually vague or contradictory institutional guidance. For some faculty, the safest move is to ban AI in the syllabus, avoid the risk, and wait for the rules to get clear. For others, the bolder move is to fold AI into their teaching quietly, using whatever tool is at hand, on their own terms, and with no one tracking what’s working or why.

Both responses are rational. Both should also raise concern. When there are no shared norms, a student’s experience with AI depends entirely on who’s teaching the course and whether their instructor embraces, bans, or ignores it. No single faculty member is to blame. It’s what happens when every instructor has to answer the same hard question on their own.

For an academic leader accountable for consistency and equity across an institution, communicating a shared standard is the first problem to solve.

No line of sight to learning

Even where AI is being used well, almost no one can prove it. And in some ways, it’s quieter and worse than inconsistency or misalignment.

Only 13% of institutions can measure whether their AI is actually working.² The rest rely on faith and anecdote, with no way to link AI use to the outcomes on which leaders are judged, such as retention, completion, and learning gains.

This is the accountability trap. An academic leader is asked to stand behind AI on campus—to a board, a cabinet, a faculty senate—without evidence of whether it’s helping students or quietly hurting them. Without evidence, a leader can’t scale what’s working, defend the investment, or improve what they can’t see.

Adoption without measurement leaves value unproven and the leader who backed it exposed.

What good actually looks like

Faculty inconsistency is a governance problem, which is the good news, because governance problems are solvable. What looks like a hundred individual choices is really one missing thing: shared norms. Setting them offers a clear, institution-wide answer to what responsible AI use looks like in teaching and learning, developed with faculty and visible to everyone.

When that answer exists, three things change.

1. Faculty stop guessing

Instructors who were frozen get permission and a path; those who were improvising get a standard to align to. Instead, every student can count on the same baseline, no matter who’s teaching.

2. Adoption gets real

Shared norms are what turn access into use. When faculty trust the institution has answered the hard questions, they bring AI into their teaching, openly. Access becomes adoption—the kind every student gets—regardless of who’s teaching the course.

3. Outcomes become visible

This is the ultimate payoff. Leaders can see whether completion rates move, whether writing improves, or whether the students who were falling behind start catching up.

That’s the through line an academic leader has to hold. Aligned faculty produce consistent use, and consistent use is the only kind you can measure, learn from, and scale. Skip the alignment and measurement is impossible. Skip the measurement and you’re back to guessing.

The leader’s move

Institution-wide AI is often framed as an IT decision or a procurement exercise. For the people accountable for what students learn, it’s a student-success decision first.

The institutions pulling ahead share a habit. Their leaders treat faculty inconsistency as something to govern. They set shared norms with their faculty, make responsible AI use the default, and build the line of sight from usage to student outcomes that so many campuses are missing.

It’s academic work, and it’s the provost’s to lead.

¹ Ellucian, “Artificial Intelligence in Higher Education: From Widespread Adoption to Strategic Integration,” Reston, VA: Ellucian, 2026, https://lp.ellucian.com/ai-industry-report.html.

² Jenay Robert, “The Impact of AI on Work in Higher Education,” research report, Boulder, CO: EDUCAUSE, January 2026, https://www.educause.edu/research/2026/the-impact-of-ai-on-work-in-higher-education.