Before lighthouses were automated, every one of them depended on a keeper. The keeper’s work was to keep the light burning in the same place every night, so that anyone at sea could find their way by it, whatever the weather.
Leaders in higher education hold that role now. AI has made the waters rough and the visibility low, and faculty, staff, and students are each steering their own course through it. A leader’s light is the institution’s mission and values, kept visible in how decisions about AI are made, shared, and revisited. When that light is steady, people can navigate with confidence even when they cannot see far ahead.
That is the leadership at the heart of Governing AI from Strength: An Action Guide on AI Governance Strategy and AI Literacies Development for Academic Leaders. D2L partnered with WCET and Opened Culture on the study behind the guide, continuing a dedication to sponsoring field-leading research that also brought the field the AI Literacies in Practice playbook.
We undertook the study because the conversation about AI governance had stayed at the level of broad strokes. Leaders everywhere agreed that governance mattered, yet few were sharing the specific practices that worked, or the lessons learned from the ones that did not. So we asked academic leaders on six continents what they were actually doing, what held, and what they would do differently.
From what they shared, we drew eight practices that make the work of guiding colleagues and stakeholders more effective, and none of them waits on a large budget. Each practice below comes with a move you can make this month, and the guide carries the evidence behind it. These practices matter most for leaders who believe they need significant resources before they can begin. What makes a lighthouse useful has little to do with how grand it is or how many keepers it has. It needs to keep shining.
1. Name Who Owns the Work
Most institutions are already governing AI. They are approving tools, writing syllabus language, and answering questions every week, often without a name for that work or anyone who owns it. The institutions that made progress named an owner first, whether a single leader, a partnership between academic affairs and information technology, or a model shared across units. AI can be everyone’s responsibility only when it is also, in a named place, someone’s.
Try this month: Name who owns AI governance and where the role sits, and say so where the campus can see it.
2. Start From What You Already Have
The leaders who moved forward did not wait to write a comprehensive AI policy. They extended the integrity, data, and acceptable use policies already on the books, and grounded new guidance in their mission and values. Guidance built that way holds up as the tools change, because it was never written for a particular tool. When an institution writes nothing, AI habits become culture by default, with none of its values built in.
Try this month: Take stock of the values, policies, groups, and people you can already build on.
3. Turn the Task Force Into a Standing Group
Task forces are a good way to start and a poor way to continue, and many lost momentum once the first questions were answered. The institutions that sustained progress planned the hand-off to a standing group with a clear charge, support between meetings, and a renewal date. A task force writes the first draft, and a standing group keeps it alive.
Try this month: If you have a task force, decide what it hands off to, and when.
4. Choose Your Pace and Put a Date on It
Leaders across the study felt the pull between moving quickly and moving carefully. The ones who moved forward chose their pace on purpose and set a date to look again, which let them act without pretending to a certainty no one has. An AI policy with a review date is a decision; one without is a hope.
Try this month: Write the date you will revisit each decision into the decision itself.
5. Begin With the AI Literacies Your People Already Hold
Good governance draws on AI literacies, the skills and mindsets people need to use AI well. Leaders already bring many of them, most often the confidence to learn by trying and the critical judgment to ask whether a use is wise. Others, such as a sense of civic responsibility, come forward when people decide together. No one person needs them all. The institutions that governed well assembled them across the table, so you can begin with the people you have.
Try this month: Look around your governance table, name the strengths already there, and find the one seat that would add what is missing.
6. Put Hands-On Experience Before a Seat at the Table
The obstacle leaders named most often was uneven understanding of AI among the people making decisions. The institutions that addressed it asked everyone in governance to complete a short preparation course, with senior leaders going first. The course asked for experience with AI and never for agreement, so skeptics kept their seats, and their informed questions made the decisions stronger. One institution put it simply: anyone may reject AI use, but no one may remain uninformed.
Try this month: Set a short preparation requirement for everyone who governs, and complete it yourself first.
7. Select Tools with Accessibility and Impact in Mind
Deciding which tools everyone can use, and who covers the cost, shapes who gets to take part. Institutions that funded a shared baseline centrally gave every department and student the same starting point, and many began with a tool they already licensed. Because tools and their terms change quickly, these institutions also scheduled regular reviews to keep the decision current.
Try this month: Record the decision, name who pays, and put the next review on the calendar.
8. Move Governance Into the Processes You Already Run
The institutions furthest along have moved AI decisions into syllabi, assessment policy, procurement, contracts, and budgeting. Governance that lives there reaches people who never attend a governance meeting, and it outlasts changes in leadership. Looking back, many leaders regretted signing contracts and deploying tools before governance caught up. A policy that no one owns is a policy that no one follows.
Try this month: Choose one process governance has not reached yet, such as procurement or syllabus review, and name who will bring it in.
Keeping the Light
None of these practices calls for a grander lighthouse. Each is part of the keeper’s ordinary work: knowing who tends the light, keeping it where people can see it, and checking it on schedule so it is ready when the weather turns. Leaders at every level can take up that work from where they stand.
For anyone who wants help putting these moves into practice, the full guide includes five worksheets and toolkits built for exactly that, from a charter for a standing governance group to a decision record for the tools everyone can use, alongside role-based actions and examples from institutions. To build the AI literacies underneath this work, AI Literacies Unlocked, the free course hosted in Brightspace, is a good place to begin. And if your institution’s governance is still in motion, the AI Literacies Case Example Database welcomes it as it stands, so others can steer by your light too.
