Prime Minister Anthony Albanese delivered a landmark address at the University of Sydney on 15 July, announcing the establishment of an Office of AI within the Department of the Prime Minister & Cabinet to develop a new, nationally coordinated approach to this rapidly evolving technology.

He promised a set of national standards for AI, allowing Australia to seize the opportunities and manage the risks. Data centres would be obliged to minimise water use and provide more energy than they consume, while Australian artists and writers would be protected from the theft of their work to train large language models.

These are important issues and a national approach is welcome, but an even bigger challenge looms for the Prime Minister’s new Office of AI – one that has so far been largely overlooked.

If we are to thrive in the age of AI – to remain, as the Prime Minister put it, a nation of innovators and inventors – we must urgently rethink our education systems. Right now, they are preparing students for a world that is vanishing.

Education and training for an AI-empowered, job-ready workforce

AI has significant implications for education and training. Our systems were designed for an era when information was scarce and cognitive labour was valuable because access to it was difficult. If AI is making these abundant, perhaps the capacity to relate deeply, deal with ambiguous situations, show care and know how to learn and listen are what our education and training systems should be developing.

But when we look at the average high school subject, undergraduate degree or training program, are these intentionally designed to foster and measure resilience? Learning how to learn? Curiosity? Empathy? The courage to lead through ambiguity?

Schools drill content because external tests reward recall over curiosity. Universities cram lectures and exams with facts. Training providers teach technical competencies. Professional bodies require content and skills while knowing that future success hinges on courage and resilience.

As the saying goes, what is measured is treasured. However, what is treasured is not always easy to measure. Of course, foundational knowledge is still key; you can’t think critically about nothing. But the system is underinvesting in the human capabilities that are becoming more valuable.

It’s getting harder to ignore the signals that education and training may be prioritising the wrong things for the AI world. Graduating students are booing commencement speakers who mention AI – they sense that the degree they just earned may already have lost its value; many employers are hiring fewer graduates into knowledge-work roles; California Governor Gavin Newsom has called for education systems to work with industry and fundamentally reimagine what they teach. We owe it to the next generation to figure this out.

Finding consensus in an AI-enabled world of study and work

As the Prime Minister said, we need to shape change together based on our shared values and standards.

It is not enough for a single university, school, TAFE provider or even state to address this alone. This is a collective action problem, and every link in the chain is currently optimised for our current needs, rather than our imminent future.

Over the last 12 months, we have been working with leaders across sectors to explore how we might reshape education and training to prepare students for an AI-influenced economy and society.

Late last year, we brought together educators, leaders, government officials, technology companies, students and professional and industry representatives for a national summit on AI in education and training. Our goal was to galvanise action to shape the future we want for learners in this country.

This cross-sector group released the Castlereagh Statement, a collective vision of how we might begin to reshape education and training, shifting the focus to what is most valuable in an AI-influenced economy and society.

What matters most in the age of AI?

Three policy opportunities emerged from this discussion and the subsequent Statement:

1. Find consensus on what we actually value in the humans we nurture in education and training;
2. Connect a disjointed lifelong learning journey across sectors; and
3. Ensure every Australian can engage with AI critically and creatively.

There is a wealth of research on the former. The World Economic Forum's Future of Jobs Report 2025 highlighted resilience, creativity, empathy, listening skills, curiosity, self-awareness and a passion for lifelong learning as key characteristics employers are looking for. The 2026 Skills Horizon report highlights building trust, unlearning, and thinking through complexity as core skills. We should build on these to develop a common understanding of which skills will be most critical in the future.

We also need practitioners, educators, employers, leaders and students to run experiments in different areas, share what works and what fails, and pilot alternatives to structures that are no longer fit for purpose. What if NAPLAN gauged resilience and curiosity in addition to literacy and numeracy?  What if school or university subjects focused on exploring wicked problems posed by community or industry, and rewarded students for failure, growth, and pushing the status quo? What if employers prioritised coachability and knowing how to learn in selection processes?

Connecting the lifelong learning journey is another challenge between largely siloed school, university and vocational training sectors, but some promising practices are emerging. The MortarCAPS group is starting to connect industry with education and training through a ‘human capability record’ that allows learners and institutions to curate evidence for core capabilities such as creativity, empathy, and adaptive leadership. The tertiary harmonisation agenda seeks to improve credit recognition and student movement, fostering cross-sector curriculum and partnerships. As society and the economy are impacted by AI, we must ensure that all pathways between education, training, and the workforce are given relevance and prestige, encouraging a flexible approach to ongoing capability development.

It is increasingly important that we teach awareness and literacy to use AI well. The OECD recently published their AI literacy framework for education, which necessarily covers much more than prompting or using AI tools. It incorporates critical awareness, human agency, partnering with AI intentionally and shaping AI systems to reflect human values. Critical and creative use of AI will include, at times, the decision to not use AI. Critically, the OECD also includes key attitudes such as curiosity, reflection, adaptability and empathy that ensure people are empowered decision-makers when it comes to how AI is used.

Our national frameworks for teaching AI skills cannot merely focus on how AI works, verifying AI outputs, building AI tools and understanding AI bias. We need to simultaneously value, build and measure the human capabilities that will define the future-ready Australian.

The Prime Minister is right that the challenges posed by AI must be addressed with coordinated national action. But alongside tackling questions of copyright and managing environmental impact, we must also take strong steps to educate every Australian for a profoundly altered world. Our productivity and global competitiveness depend upon it.

Danny Liu is Professor of Educational Technologies at the University of Sydney.

Jason Lodge is Professor of Educational Psychology at the University of Queensland.

They were two of the conveners of a national summit that brought together experts from across Australia's education and training sectors to develop the Castlereagh Statement.

Image credit: iStock #2244328282