How is your institution holding the AI policy & guidance tension?

[ZC]

Maybe you've seen today's post on the Google Keyword Blog reflecting on our pilot AI Policy & Guidance Labs conducted globally over the winter. 

As we were putting that post together, and as we were building and running the labs, I had a key point from the Educause 2025 conference running through my mind. In more than one session, as Institutions of Higher Education (IHEs) were talking about their AI journeys, they spoke to data from their faculties pointing to a key tension - the want for institutional guidance but not so much that it invades faculty autonomy.

I started looking into this more after I got home from Nashville and found it was a tension uncovered across the field. 

The report on the results of the 2025 WCET survey, Supporting Governance, Operations, and Instruction and Learning Through Artificial Intelligence put it this way, "Responses suggested a possible tension between faculty autonomy in developing and implementing classroom policy, and the need for adoption of more holistic, campus-wide policies and guidelines."

Further, this pre-print of Faculty Readiness for AI-Supported Teaching and Scalable Online Program Delivery in Higher Education... frames one aspect of the problem thusly:

The salience of guidelines is reinforced by the inverse: 80% of faculty do not find institutional AI guidelines comprehensive, suggesting that policy ambiguity functions as a readiness inhibitor (Digital Education Council, 2025).

Institutional leader evidence further demonstrates that governance maturity is still developing. CHLOE reports that 35% of institutions have institution-wide AI policies, while 40% are still discussing but have not published policies, implying that many faculty are operating in policy vacuums or fragmented departmental regimes (Simunich et al., 2024). Complementarily, EDUCAUSE frames AI policy development as a multi-domain institutional project spanning governance, operations (including professional development and infrastructure), and pedagogy (including integrity and assessment) (Robert & McCormack, 2024)...

The evidence indicates that readiness is structurally constrained by the institution’s support ecology. If policy clarity and training are missing, faculty may rationally restrict student AI use (as reflected by high prohibition rates) to preserve assessment credibility (Robert & McCormack, 2024; Ruediger et al., 2024). From an organizational readiness lens, policy immaturity and insufficient resources reduce collective change efficacy, lowering implementation quality even where change commitment exists among some faculty (Weiner, 2009).

Across several other sources listed below, one read of the picture IHE faculty and their leadership face is how to right-size the professional learning, policy, and guidance from an institutional level while also meeting faculty needs. 

IHEs are also faced with the fact that the use of these technologies doesn't play by the same rules as previous "disruptions." Learning management systems, multimedia learning objects, wikis, blogs, social media platforms, etc. Each is a tool allowing instructional faculty to operate at some level as a gatekeeper to their use within a course. If I eschewed my institution's LMS in favor of a Google Site, that was my choice as an instructor.  It was not likely my students would decide to build their own MOODLE install and share it with their classmates. AI-enabled tools make such gatekeeping much more difficult if not impossible. Their existence is a forcing function to shifting practice.

Some lock everything down. Blue books and #2 pencils. Others play whack-a-mole, shifting policies in the moment, unsure how to re-design assignments to meet the moment. A few recognize the need for malleability, instituting systems like the AI Assessment Scale to build a common language as they continue learning and designing this new knowledge landscape.

What about you? 

How are you and your institution navigating the moment?

How are you building policies and guidance that hold learning at the center while navigating the flow of new tools and capabilities? 

Further Readings:

Procedure to Evaluate Artificial Intelligence (AI) Output for Educational and Clinical Tasks

Empowerment vs. Oversight: Student Perspectives on Classroom GenAI Policies

The AI Implementation Gap in Higher Education: Navigating the Disconnect Between Technology Adoption...

[relwell]

I shifted from the classroom to school to system K-12 work, and now have gone into Tertiary. The positive is that it is a smaller universe to control, so there is at least room for conversations, whereas the system level stuff was a non-starter. The main point of friction, and endless annoyance, is that edu-focused teams and faculties are told to be bold and innovative and forward thinking. But, the ICT teams that control access to the things like AI are on a highly constrained, hyper-risk-averse setting.

I am glad to say my university, while not embracing, is certainly not avoiding AI. We just released assessment classification for AI use info for staff that makes it mandatory to address AI use on tasks before they go out into the wild, and we've avoided any endorsement of using tools like Turn it in for the void of hopelessness that is AI Detection.

We are not a Google institution yet, but part of bringing me over was my digital transformation experience at K-12 where I got Gemini across the line for our teachers. We're running a pilot in my winter sem courses, and I have the momentum (I hope) to not only enable GfE for teachers and staff in our Faculty of Education, but also help drive admin console work to ensure that pedagogy and teacher/student safety are the main focus for what we enable and how we message it. Just means I have to spin up a PL program for staff, but that is my wheelhouse so I am confident there, at least!

Parting shot - I presented at Moodle Moot in Sydney last year, and was surprised at the number of highly regarded institutions adopting a 'wait and see' stance. It made me uncomfortable then, and more so now. I just don't see that being a winning strategy regardless of how things play out.