When generative artificial intelligence tools such as ChatGPT became publicly available in late 2022, their impact on education was immediate. Students began experimenting with AI generated essays, teachers explored automated lesson planning, and school systems were confronted with an uncomfortable reality. The technology had arrived before policy was ready.
A new peer reviewed study led by Professor Matt Bower of Macquarie University sheds light on how senior education policymakers are responding to this disruption. Published in The Australian Educational Researcher, the article titled “What generative Artificial Intelligence priorities and challenges do senior Australian educational policy makers identify and why?” offers rare insight into the thinking of those tasked with governing AI use in schools.
Rather than focusing on speculative futures, the research captures a critical moment in real time. It examines how Australian education leaders attempted to develop ethical and responsible policy while grappling with urgency, uncertainty, and the rapid pace of technological change.
Why policymakers matter in the AI in education debate
Much of the public conversation around artificial intelligence in education has focused on teachers and students. Yet policy decisions made behind closed doors determine whether AI becomes a tool for educational innovation or a source of systemic risk.
Senior policymakers operate at the intersection of ethics, governance, pedagogy, and public trust. Their decisions influence curriculum reform, assessment standards, professional development, data governance, and equitable access to technology across entire education systems.
Professor Bower and colleagues argue that understanding policymaker priorities is essential. Without this insight, debates about AI in schools risk being driven by media narratives or commercial interests rather than evidence informed governance.
Inside the study and who was involved
The study was conducted by researchers from Macquarie University, Monash University, the University of Queensland, and the University of Sydney. Data were collected during a pivotal period between early 2023 and late 2023, before Australia released its national framework for generative AI in schools.
Twenty two senior education leaders participated, representing state and federal education departments, school sectors, teacher unions, national education agencies, and digital education providers. Many were directly involved in shaping national AI policy.
Using a mixed methods qualitative design, the researchers combined surveys, ranking exercises, and in depth focus group discussions. This approach allowed participants to identify priorities, debate risks, and explain the reasoning behind their decisions.
Managing risk emerged as the top priority
Across the study, one concern consistently rose to the surface. Risk management.
Participants ranked managing risks such as misinformation, data privacy breaches, biased outputs, and academic integrity as their highest priority. The concern was not abstract. Policymakers repeatedly emphasised that AI tools were already being used in classrooms with little guidance or oversight.
Several described the situation as an emergency response rather than a measured rollout. Generative AI systems are designed to produce confident answers, even when those answers are inaccurate. In an educational context, this raises concerns about student learning, trust in assessment, and the spread of false information.
Teachers are central to every policy decision
While risk management topped the list, educating teachers ranked a close second. Policymakers repeatedly stressed that no AI policy can succeed without teacher capability and confidence.
Teachers are the primary translators of policy into classroom practice. They decide how AI tools are used, how student work is assessed, and how ethical boundaries are enforced. Yet many teachers have limited training in artificial intelligence, particularly generative models.
The study highlights a tension facing policymakers. On one hand, there is pressure to act quickly. On the other, teachers are already experiencing workload strain and professional fatigue following years of pandemic related disruption.
Participants emphasised the need to build deep conceptual understanding rather than platform specific training. Given the rapid evolution of AI tools, professional learning focused on how a single product works risks becoming obsolete within months.
The pace of change is overwhelming the system
Among all systemic and environmental challenges identified in the study, one stood out with near unanimous agreement. The pace of change.
Every participant ranked the speed of AI development as a major challenge. Policymakers described AI disruption as arriving on top of existing pressures, including post pandemic recovery, teacher shortages, curriculum reform, and digital transformation initiatives.
This accelerated pace complicates evidence based policymaking. Education systems are traditionally cautious and deliberative, yet generative AI is evolving at a speed that outstrips legislative and regulatory processes.
Participants acknowledged that policy is often reactive rather than anticipatory. In many cases, schools and students adopt technologies before governments can assess risks or establish safeguards.
Equity and access remain unresolved challenges
Another recurring theme was concern about equitable access to generative AI technologies. Policymakers warned that AI could exacerbate existing educational inequalities if access is determined by school funding or household income.
Advanced AI tools are increasingly available through paid subscriptions, integrated productivity platforms, or high performance devices. Schools in low socioeconomic or remote communities may lack the infrastructure required to benefit from these technologies.
Participants feared the emergence of a two tiered education system. In such a system, well resourced schools gain access to sophisticated AI support while disadvantaged students are left behind.
Equity concerns also extended beyond access to include representation in training data, cultural bias in AI outputs, and the potential marginalisation of Indigenous and minority perspectives.
Assessment integrity is under pressure
High stakes assessment emerged as a particularly sensitive issue. Policymakers expressed concern that generative AI could undermine confidence in qualifications if students submit AI generated work without demonstrating genuine learning.
This issue is not limited to plagiarism. Generative AI challenges traditional notions of authorship, originality, and cognitive effort. Policymakers noted that assessment practices designed for a pre AI era may no longer be fit for purpose.
Some participants viewed this disruption as an opportunity. AI could act as a catalyst for long overdue assessment reform, encouraging more authentic, process oriented evaluation of student learning.
However, redesigning assessment at scale requires time, expertise, and political will. In the interim, uncertainty persists.
Policy making in an environment of uncertainty
Beyond specific priorities and challenges, the study reveals how policymakers experience AI governance itself. The researchers identified five meta themes that characterised the policy environment.
Urgency was driven by widespread AI adoption before policy frameworks were in place. Uncertainty reflected limited empirical evidence about educational impacts. Interconnectedness described how decisions about AI affect curriculum, assessment, equity, and teacher professionalism simultaneously.
Contextuality acknowledged that schools differ widely in resources and readiness. Complexity captured the difficulty of balancing innovation with risk, and speed with deliberation.
Australia as a case study with global relevance
Although the research focuses on Australia, its findings resonate internationally. Education systems worldwide are confronting similar challenges as generative AI tools become embedded in everyday digital platforms.
The absence of comprehensive AI regulation at the time of ChatGPT’s release is not unique to Australia. Policymakers globally are struggling to align national legislation, ethical frameworks, and education specific guidance.
The study illustrates how governance gaps can place pressure on schools and teachers to make ad hoc decisions. It also highlights the importance of national collaboration in avoiding fragmented or inconsistent policy responses.
Reference
Bower, M., Henderson, M., Slade, C., Southgate, E., Gulson, K., and Lodge, J. (2025). What generative artificial intelligence priorities and challenges do senior Australian educational policy makers identify and why. The Australian Educational Researcher, 52, 2069 to 2094. https://doi.org/10.1007/s13384-025-00801-z
