The honest answer is that no one fully is. It is a feature of a particular kind of challenge that resists being handed to the right department or solved by the right hire.

Two types of problems

Ronald Heifetz and his colleagues draw a distinction I keep returning to: the difference between technical challenges and adaptive ones (Heifetz et al., 2009).

Technical challenges are problems where existing expertise applies. They may be difficult or expensive, but the path forward is knowable. Migrating to a new learning management system is a technical challenge. Deploying a plagiarism detection tool is a technical challenge. The solution exists; it is a matter of implementation by the people with the right expertise.

Adaptive challenges are different. They require people who experience the challenge themselves to change how they think, what they value and how they behave. There is no clear pre-existing solution to reach for. The problem itself is often unclear until you are well into trying to address it. And critically, the discomfort of working through it cannot be outsourced.

Generative AI in higher education is, at its core, an adaptive challenge. Not just because these tools are novel and unfamiliar, but because they surface genuinely unsettled questions: What does it mean to learn something? What is the point of a written assignment? What does a degree certify that AI cannot replicate? These are not questions teachers, educational developers, IT or HR can answer fully — not even a vice-rector can answer them on behalf of everyone else.

Leadership as a practice, not a position

This is where I want to push back on how “leadership” tends to get used in these conversations. It usually implies someone senior: a rector, a dean, a director of education. Those people carry real responsibility for the conditions in which change happens. But adaptive challenges are not solved from the top.

Marshall (2010) argued that university culture and existing capability significantly constrain organisational change — that in the absence of strong direction, technologies tend to get absorbed into whatever is already happening rather than driving genuine transformation. This is worth sitting with. It means that even when an institution responds, convening the working group, issuing the policy, commissioning the report, the default trajectory is assimilation. The new thing gets shaped by existing practice rather than reshaping it.

That observation cuts both ways, though. The same distributed authority and professional autonomy that slows coordinated response also means that genuine change tends to originate from unexpected places: from people closest to the problem who have the most at stake in how it resolves. Heifetz would probably call this mobilising the system. Kowch (2021) makes a related argument: formal leadership structures can actually limit innovation by over-organising authority into vertical hierarchies, and adaptive change requires less formal, more networked forms where innovation can surface from wherever it emerges.

What that means practically is that leadership, in this context, is less about authority and more about a willingness to raise uncomfortable questions, hold the tension of not-yet-having-answers and keep the conversation focused on what actually matters rather than on the nearest technical fix. Kouzes and Posner (2019) frame it precisely this way: leadership is a set of behaviours (ways of modelling, challenging, enabling and encouraging) that anyone can practise regardless of where they sit in a hierarchy. The lecturer rethinking her assessment, the programme coordinator trying to articulate a coherent position for his team, the student advisor fielding questions she has no institutional guidance on yet. All of them are doing leadership work, whether or not it features in their role description.

The temptation of the technical fix

What makes adaptive challenges genuinely hard has less to do with their complexity than with the pull towards treating them as technical problems. If your house is too cold, you turn up the thermostat. But if the problem is that you built your house for a climate that no longer exists, adjusting the thermostat might offer some comfort without resolution. You are managing a symptom while the situation remains unchanged.

Much of the early institutional response to GenAI followed exactly this pattern: detection software, integrity policies, one-off workshops, each addressing the visible surface of the problem while leaving the harder questions untouched. Most responses still do. They are really about coping with the tools in the now: what to allow, what to detect, which chatbots to adopt. What tends not to appear is the harder, prior question: what education is actually for and what must not be lost in the process of adapting to this moment.

The Castlereagh Statement (Castlereagh Summit, 2026), a cross-sector response to GenAI in education built by more than eighty practitioners from across Australia, is worth reading precisely because it resists this move. Where most institutional responses reach for a fix, it names the bigger picture, complexities, questions and values at stake, such as:

Education is fundamentally relational. What matters is not the evaluation of outputs but the understanding and connection through which learning actually happens.

That framing is rarer than it should be.

The step that keeps getting skipped

A plan is not the problem. The problem is what the plan tends to be about. Institutional responses to GenAI are rarely short of structure: working groups, policy timelines, procurement decisions, training calendars. What they are often short of is a plan that takes the adaptive dimension seriously, one focused on what needs to genuinely change in how people think, work, teach and learn, rather than what needs to be installed, detected or permitted.

What gets skipped, at every level, is the step that requires the most patience. The one about people, roles, values, sensemaking. That step determines whether everything else actually lands, or hardens into practice built on beliefs about teaching and learning that were never made explicit, let alone shared.

It also requires that the people who encounter this challenge most directly, those teaching through it, advising through it, designing curriculum through it, are given the time and space to work through it together. Not as a one-off workshop. As a genuine, protected part of how an institution thinks.

That work belongs to everyone in the room. Not just the person who called the meeting.

References & further reading

Castlereagh Summit (2026) The Castlereagh Statement: A cross-sector call to action on Australian
education and training in the age of AI. Sydney, Australia. https://castlereagh.ai/

Heifetz, R. A., Grashow, A., & Linsky, M. (2009). The practice of adaptive leadership: Tools and tactics for changing your organization and the world. Harvard Business Press.

Kowch, E. (2021). Leading transformation with digital innovations in schools and universities: Beyond adoption. In D. Ifenthaler, S. Hofhues, M. Egloffstein, & C. Helbig (Eds.), Digital transformation of learning organizations (pp. 145–168). Springer. https://doi.org/10.1007/978-3-030-55878-9_9

Kouzes, J., & Posner, B. (2019). Leadership in higher education: Practices that make a difference. Berrett-Koehler Publishers.

Marshall, S. (2010). Change, technology and higher education: Are universities capable of organisational change? Australasian Journal of Educational Technology, 26(8), 179–192. https://doi.org/10.14742/ajet.1018