Beyond the Framework: What Participants Said AI Implementation Requires

Brett Christie, Ph.D.
VP, Educational Innovation & Inclusivity

When nearly 200 participants gathered on July 22, 2026, for our AI on Campus webinar “Beyond the AI Framework: Adoption, Adaptation, and Lessons from MUSC,” the focus was not whether institutions should provide AI guidance. That question has already moved to the background for many colleges and universities. The more urgent question is what happens after guidance is written, shared, and officially endorsed.

The session featured Dr. Julaine Fowlin, E’lise Nissen, and Dr. Matthew Greseth from the Medical University of South Carolina (MUSC), who returned one year after sharing MUSC’s AI Acceptable Use Framework for Academic Tasks. This follow-up session examined how an institutional framework moves into practice through policy, faculty development, course redesign, student guidance, and ongoing adaptation.

During the webinar, participants were invited to respond to two Padlet prompts:

Most difficult part of moving AI guidance from institutional document into teaching and academic practice?

Which strategy, resource, or practice has helped your institution make progress?

Taken together, their responses provide a useful snapshot of where institutions are in the work of academic AI implementation.

Prompt 1: What Makes AI Guidance Difficult to Move Into Practice?

1. The challenge of vision, leadership, and sustained ownership

Several responses pointed to the need for clearer institutional direction and sustained leadership. Participants named “lack of vision,” “lack of clear direction from leadership,” and the question of “who leads it that will stick around and keep momentum?” These comments suggest that AI implementation can stall when responsibility is diffuse or temporary.

The issue is not only whether an institution has guidance. It is whether there is an identifiable structure for carrying the work forward. Participants seemed to be describing a gap between publishing an institutional position and creating the conditions for ongoing adoption, revision, communication, and accountability.

2. Faculty, administrator, and student buy-in remain central

A second theme was buy-in across stakeholder groups. Participants named faculty buy-in, administrator and faculty buy-in, student buy-in, and the difficulty of “getting people to pay attention.” One response described the challenge as “teaching students and faculty to develop a positive AI mindset.”

These responses suggest that AI guidance is not implemented simply by making it available. It must compete for attention, address skepticism, and engage deeply held concerns about academic quality and student work. One participant captured this concern directly by naming the difficulty of “trusting that learning will still occur.”

3. Time and planning are major implementation constraints

Participants repeatedly pointed to time as a barrier. One response named “time, shared agreements, lack of knowledge,” while another described a policy approved in May with expectations for a fall rollout, noting that the timeline did not align with the amount of work required.

This theme is important because AI implementation often appears, from the outside, to be a policy or communication task. The Padlet responses suggest something different. Moving guidance into practice requires time for curriculum updates, faculty consultation, student communication, assignment redesign, and consensus building. Institutions may underestimate the implementation labor that begins after a policy or framework is approved.

4. Decentralized teaching makes consistency difficult

Another theme was the decentralized nature of academic work. One participant noted that faculty have ideas but are often “working in silos,” with a workgroup only recently created. Others named the challenge of stakeholder mobilization and shared agreements.

This reflects a common institutional tension. Teaching decisions often live at the course, instructor, department, or program level, while AI guidance is often written at the institutional level. The implementation challenge is connecting those layers without flattening disciplinary judgment or faculty agency.

5. Institutions are trying to move from binary rules to nuanced practice

One especially useful response described the challenge as “such a wide range” of thoughts, feelings, and approaches to AI among faculty, staff, and students. The participant noted the difficulty of moving from binary thinking toward a more nuanced framework such as the AI Assessment Scale.

This captures a key shift in the field. Many early conversations framed AI use as allowed or prohibited. Participants’ responses suggest that institutions now need more precise ways to discuss when AI use supports the task, when it undermines the task, and how expectations can be made clear to students.

6. Practical concerns remain unresolved: privacy, access, tools, and testing

Some responses surfaced operational and ethical complications, including data privacy concerns, blocked AI resources for students, and the need for user testing with intended audiences. Others gestured toward tool-related limits and the time required to test guidance, resources, or assignments before broad adoption.

These responses remind us that AI implementation is not only pedagogical. It is also technical, legal, ethical, and infrastructural. Faculty guidance can falter if students do not have access to approved tools, if privacy expectations are unclear, or if resources are rolled out before they have been tested in actual teaching contexts.

Prompt 2: What Helps Institutions Make Progress?

1. Shared frameworks and language help move the conversation forward

The AI Assessment Scale (Perkins, et al., 2024) appeared prominently among the helpful resources named by participants. One response noted that AIAS had supported individual faculty adoption and was moving toward broader faculty-level adoption. This suggests that faculty need more than general principles. They benefit from shared language that can be applied at the level of assignments, courses, and student expectations.

A common framework can also reduce the burden on individual instructors. Instead of each faculty member inventing their own AI categories and terminology, a shared scale gives institutions a starting point for communication, adaptation, and course-level decision-making.

2. Committees and workgroups provide structure for coordination

Participants also named the value of organizing a committee. In light of the barriers named in the first prompt, this is significant. Committees, workgroups, and cross-functional teams can help institutions move beyond isolated experimentation.

The most useful groups are likely those that connect academic leadership, faculty, faculty development, instructional design, educational technology, student support, and policy stakeholders. Without that coordination, AI guidance may remain fragmented across units or disconnected from faculty practice.

3. Faculty development must be practical, ongoing, and supported

Several responses pointed to faculty development mechanisms, including CTL short courses, preliminary training on licensed AI tools, panel discussions, and communities of practice. These responses reinforce that implementation depends on repeated opportunities for faculty to examine examples, ask questions, revise materials, and compare approaches with colleagues.

One participant suggested giving faculty a course release, which signals an important point: meaningful AI integration takes time. If institutions expect faculty to redesign assignments, clarify expectations, revise syllabi, and engage students in new conversations about AI use, that work requires capacity, not only encouragement.

4. Communities of practice and learning communities reduce isolation

Participants named communities of practice and learning communities as helpful strategies. This directly responds to the challenge of faculty working in silos. Shared spaces allow faculty to test ideas, see examples from peers, and normalize the reality that AI guidance will continue to evolve.

One response described small-group workshops and learning communities where participants created ready-to-use widgets using “vibe coding.” Whether focused on course materials, assignment supports, or AI-enabled tools, the larger point is that faculty progress often happens through collaborative making, not just passive training.

5. Examples and student-facing resources make guidance more usable

Participants also pointed to examples, support when students complain, and iterative student and faculty resources. These responses suggest that implementation depends on concrete materials that help faculty and students interpret guidance in context.

This is especially important when AI expectations vary by assignment. Faculty need examples of syllabus language, assignment instructions, disclosure statements, and AI-use categories. Students need guidance that helps them see what is permitted, what must be documented, and why expectations may vary across courses or tasks.

6. Action matters more than perfect readiness

One participant captured a broader implementation lesson with the phrase “Analysis Paralysis,” followed by the comment, “Inspiration is good, action may be better.” This theme is worth highlighting because many institutions are still trying to produce definitive guidance in a rapidly changing environment.

The Padlet responses suggest that progress often comes through iteration: creating a committee, piloting a framework, offering short courses, testing examples, gathering feedback, and revising resources over time. Institutions may not be able to wait for perfect clarity before helping faculty and students navigate AI use in academic work.

What the Responses Reveal About the State of AI Implementation

Across both prompts, participants described institutions that are no longer simply debating whether AI guidance is needed. Many are now confronting the harder work of implementation: building buy-in, creating shared language, supporting faculty redesign, addressing student transparency, coordinating across decentralized structures, and keeping guidance current as tools change.

The most consistent message was that AI guidance does not become institutional practice on its own. It needs leadership, time, faculty development, student-facing resources, cross-functional coordination, and a willingness to revise. A framework may be the visible artifact, but the deeper work is cultural and operational.

In that sense, the participant responses closely echoed the central theme of the MUSC session. The work “beyond the framework” is not simply about adoption. It is about adaptation, capacity building, and sustained institutional learning.