The share of institutions with an AI acceptable use policy nearly doubled between 2024 and 2025. The institutions that made that leap are ahead of the field on governance. Yet most of them are still stuck on implementation.
That gap between policy and practice is where AI momentum stalls, even at institutions that moved quickly and ambitiously to embrace it. Closing it starts with understanding where it opens.
Faculty experimentation is outrunning governance
According to the 2025 EDUCAUSE AI Landscape Study, faculty training ranked as the most common element in institutions' AI strategic planning efforts, and teaching and learning led all functional areas in AI focus. Faculty are on the front lines of AI in higher education, adapting courses, testing tools, and making judgment calls about usage, often without a shared framework to guide those decisions.
When faculty innovate without shared frameworks, the results are uneven. One instructor builds AI-supported course materials with thoughtful pedagogical scaffolding. Another uses it to generate a module outline without checking alignment to learning objectives. Both show up as AI adoption in a survey; only the first builds toward anything durable.
And the policies that did get adopted still cover a minority of institutions. Even where one exists, a policy doesn't tell faculty how to redesign a course. Governance frameworks built alongside faculty development give all of that individual innovation somewhere to land.
The instructional layer has no owner
Institutions get stuck between strategy and practice for a consistent reason: the work of translating institutional AI direction into course design has no clear owner in most AI strategies.
According to a 2026 Government Technology analysis of AAC&U research, the institutions making the most progress blended top-down and bottom-up approaches, and what most of them shared was treating course design as part of the strategy rather than something to sort out after it was set.
That distinction has consequences. A 2025 CITE Journal systematic review found that instructional designers are increasingly expected to bridge policy and course-level practice, handling everything from assessment redesign to tool protocols while coordinating across faculty, IT, and compliance teams, often without the authority, resourcing, or policy guidance the work requires. The institutions making headway have put a name and a budget line next to that work.
Policies are arriving before the practice they govern
A policy document takes months to write. Instructional practice takes years to build, and the distance between the two is wider than most institutions currently acknowledge. EDUCAUSE survey data from 2025 found that more than half of institutions were already using AI to support curriculum design and administrative workflows. For most of them, the instructional frameworks to govern that use came second, if they came at all.
So a well-crafted institutional AI policy can sit in a faculty handbook while course design continues exactly as before. Faculty aren't resisting it; in most cases, no one has translated the policy into the decisions a course designer faces on a Tuesday.
The institutions making the most progress run policy development and instructional practice development side by side instead of in sequence.
Course design is where this has to get resolved
Policy sets the boundaries of what's permitted; the learning itself takes shape at the course level, and the two need to be in close conversation. A faculty member who receives an institutional AI policy and a list of approved tools still faces the harder questions: how to sequence a module so AI supports rather than shortcuts student thinking, how to design activities that develop the skills the course is meant to build, and how to know when AI adds rigor versus when it removes it. These are pedagogical questions, and they require instructional expertise in addition to policy compliance.
EDUCAUSE's 2026 research on AI and work in higher education found that only 15% of higher education professionals reported their institutional leaders as cautious or very cautious about AI. Leadership readiness, in other words, has largely arrived.
Whether AI adoption becomes embedded practice or stays a collection of individual experiments depends on how well institutions bridge the space between policy and the course designer's day-to-day work.
Instructional design gets a seat before the strategy is set
Look across the institutions closing the gap and the common thread is where instructional design sits: inside the strategy conversation before the strategy is finished, treated as a design constraint from the start rather than an implementation detail for later. A more sophisticated AI policy helps. A larger technology investment helps. Neither one substitutes for that seat at the table.
These institutions also resource the role before the strategy ships, with the authority and budget the work requires, instead of handing it down to instructional designers already stretched across assessment redesign, tool protocols, and cross-functional coordination.
Academic leaders who treat course design as a downstream concern will find that every other AI investment lands somewhere upstream of where it needs to go.
If your AI strategy is in place and the course design layer is where momentum stalls, that gap is exactly what the SRM AI + Human Insight framework was built for: the practice layer, where instructional decisions get made. The AI+HI Hub is where to start.
Sources
- EDUCAUSE. (2026). The Impact of AI on Work in Higher Education. educause.edu
- Robert, J. (2025). 2026 EDUCAUSE Top 10: Making Connections. EDUCAUSE Review. er.educause.edu
- EdTech Magazine. (2026). AI Governance in Higher Education: An Overview. edtechmagazine.com
- Sourwine, A. (2026). Implementing AI for Higher Ed: From the Top Down, or Bottom Up? Government Technology. govtech.com
- CITE Journal. (2025). AI-Integrated Instructional Design in Higher Education: A Systematic Exploration of Tools, Roles, and Challenges. citejournal.org
- Sourwine, A. (2025). Survey: Higher Ed AI Adoption Faces Financial, Policy Hurdles. Government Technology. govtech.com
