This blog presents a version of an Adult Learning Australia Webinar that my colleague (Dr Garry Argent) and I presented on Thursday, June 4, 2026.
It takes the form of a conversation (in 3 Acts) in which Garry and I take turns to explore key aspects of a collaboration spanning more than 10 years. That collaboration produced Garry’s PhD thesis, and a jointly authored book where we built on that PhD work to explore emerging challenges in relation to Language Literacy and Numeracy (LLN) skills as the Age of Machine Intelligences promises to transform what we understand by those skills, and the worlds of work in which those skills might be deployed.
We introduced this previous work in an earlier blog – https://youngpeoplesfutureslab.org/new-book-on-the-problem-of-foundation-skills-and-the-future-of-work/
Introduction
Garry: I want to start with a student.
His name was Craig — that’s a pseudonym — and he was a foundation skills learner in an AMEP I was teaching. He’d recently discovered ChatGPT, and he was using it to write emails for his small ICT and mobile phone business outside of class. When I gave the Level 3 class a writing assessment, Craig asked whether he could use ChatGPT to complete it. He was enthusiastic about it — he told me:
‘It’s so fast, I can ask it to write an email and it’s done!
Or sometimes I write the email and it fixes it for me.
It’s good for my business, I can’t have bad mistakes.’
I, fearing the ‘new’ and the ‘unknown’, responded reflexively with concerns about plagiarism and the assessment’s objective of the students demonstrating their knowledge of the writing process. Craig, not one to argue, accepted my hastily cobbled together response.
But as he moved on, he said something I couldn’t quite let go of:
‘mmm… maybe one day people don’t need to learn writing because of ChatGPT.’
I didn’t have a good answer. I’m still not sure I do. But that moment is where this book begins — and it’s where we want to begin today. We’ll come back to Craig.
Peter: What draws Garry and I together in this work is a shared unease about a gap — the gap between how foundation skills policy imagines learners, and who those learners actually are. Between what the system says it’s doing, and what it is doing.
Garry spent years working in Learn Locals, community and industry RTOs, and in TAFE — much of it in adult literacy, and in migrant and refugee programs including AMEP and SEE. What he kept encountering were people navigating work and life without the skills the system assumes everyone already has. And a policy architecture that claimed to help them, but was built on assumptions that didn’t quite fit.
My own background is in researching the lives of marginalised, disengaged and historically disadvantaged young people — their education, their employment, their health and sense of belonging — in the context of what I’d call the compounding crises of our moment: economic precarity, climate disruption, and now the rapid emergence of machine intelligences that are reshaping what work and learning mean.
The book that came out of Garry’s PhD brought those two perspectives into conversation. And today we want to take you through three connected ideas from that work.
First — where the current foundation skills apparatus came from, and what kind of logic built it.
Second — what that apparatus does to learners: specifically, how it positions them as choice-making consumers in a market that doesn’t always behave like one.
And third — where foundation skills might be heading, in a world that Craig’s question already anticipated.
We’re not going to resolve these questions today. But we hope to introduce you to a sharper set of tools for thinking about them — and we want to hear what you’re thinking too. There’ll be space for that as we go, and a fuller discussion at the end.
Let’s start with the history.
Act 1: Where Did This Come From? A History of the Foundation Skills Apparatus.
Peter: Before we get into the history, I want to touch on two ideas that do most of the work in this part of the book.
The first is the idea of the neoliberal. We use this word carefully, because it gets thrown around a lot. What we mean by it, in this context, is a particular way of governing — individuals, systems, organisations — that works through three connected moves.
The first move is privatisation: services that were once delivered by governments — training, employment support, skills programs — are transferred, gradually or suddenly, to private agencies, both for-profit and not-for-profit.
The second move is commodification: once something is privatised, it starts to behave like a commodity — something that can be bought, sold, and competed over. Foundation skills development becomes a product. Learners become customers.
The third move is individualisation: individuals are increasingly encouraged — through policy, through program design, through the language of choice and responsibility — to think of themselves as autonomous agents, navigating these markets on their own behalf, responsible for their own outcomes.
The second idea is apparatus. By this we mean the architecture that gets built when governments decide something is an urgent problem needing a response. An apparatus includes 3 things:
- the way the problem gets defined — because how you name a problem shapes what solutions become possible;
- the agencies, policies, funding streams and regulatory bodies that get created or mobilised to address it;
- and the relationships — between providers, governments, learners, regulators — that those structures then require.
You can’t see an apparatus. You can’t point to it. But you can trace it — through documents, through policy decisions, through what gets said and, just as importantly, through what gets left out.
That’s what we did, and Garry, is going to take us through some of what we found
Garry:
What we found, when we went back through the archives, was that the foundation skills apparatus we have today didn’t arrive fully formed. It was built — piece by piece, decision by decision — and understanding where it came from helps explain why it works the way it does.
Most of the documents we examined weren’t even about foundation skills directly. But as we traced the processes of privatisation, commodification and individualisation that Peter has just outlined, we kept asking: why did these particular moves happen when they did? What did they make possible — and what did they silence?
Let me take you through three moments that shaped what we have today.
The first moment: the Hilmer Review, 1993.
Competition policy didn’t emerge from nowhere. In the early 1990s, there was a strong sense that Australia needed to become more productive and more competitive in a changing global economy. The Hilmer Review was the response — and what it did was embed competition as a central organising logic, not just for markets, but for social systems as well.
For training, the practical consequence was significant. Rather than funding being allocated directly to providers, Commonwealth and state programs began to be put out to competitive tender. TAFEs, private RTOs, community providers — all of them now competing for funding, contracts, and students.
If you’re in this room, you probably remember when that shift started to feel real in your own organisation.
The second moment: the OECD Adult Literacy Surveys — 1996, 2006, 2012.
These surveys gave governments across the OECD a number — and numbers, in policy, are powerful. The finding that large proportions of working-age adults lacked the literacy skills needed for contemporary work and life produced reactions that were, by any measure, dramatic.
In North America, the language was almost panicked:
‘Half of America’s adults are functionally illiterate!’
‘People are illiterate, and they don’t even know it! This is very dangerous!’
In Australia, the surveys connected with an already well-established focus on skills as human capital — the idea that literacy deficits were a drag on productivity and economic competitiveness. With just under half the population estimated to lack the foundation skills needed for work and life, the “problem” was now statistically certified.
For example the Adult Literacy and Life Skills (ALLS) survey in 2006:
43.5 per cent (approximately 6 million) of working age Australians (aged 15-64) had not attained skill Level 3 in document literacy and 49.8 per cent (approximately 6.8 million) had not attained skill Level 3 in numeracy…. the minimum level required by individuals to meet the complex demands of work and life in modern economies.
This data was drawn upon repeatedly in policy, policy responses and research.
The third moment: the Victorian Training Guarantee, 2008 to 2012.
This is the moment that will feel most familiar to many of you — because its consequences are still being felt.
The Victorian Government introduced what was, in effect, a fully marketised training system. The promise was flexibility. Providers would be able to deliver subsidised training in response to demand rather than fixed allocation — customising programs, responding to individual and business needs, offering specialisations. The language was explicitly about consumer choice and market responsiveness.
As outlined by DIIRD (2008, p.23), providers “will be able to deliver government-subsidised training in response to demand rather than by fixed allocation. More providers will be able to customise training products to the needs of individuals and businesses, and offer specialisations that reflect their own expertise.”
What happened instead was rapid expansion, budget blowouts, and a system that became, in parts, genuinely abused. Aggressive marketing. Low- or no-cost courses as inducements. Students enrolled in multiple courses without their knowledge or understanding. And — and this is directly relevant to foundation skills — significant mismatches in placement: learners with very low LLN skills enrolled in Diploma-level courses; learners with postgraduate qualifications placed into foundation skills units.
The response from government was not to reconsider the marketised model. It was to regulate it more tightly — to protect consumer confidence, manage reputational risk, and reinforce the system’s status.
Victoria therefore became a key reference point for other states and territories — both as a model ‘not’ to follow, and as a cautionary example in reforming their own vocational education systems.
Or as Leesa Wheelahan describes in her commentary a ‘great template in how not to reform vocational training’.
It was in this environment that a national regulator — ASQA — was established.
And it was this environment that shaped the Commonwealth programs many of you work in or alongside, including AMEP and SEE.
Peter: What we want to draw out from these three moments is something that sits underneath all of them — a shift in who is responsible.
Under the earlier model, government was responsible for provision. It funded and delivered training as a public good.
Under the apparatus that these moments built, responsibility begins to shift. To providers, who must now compete to survive. And — as we’ll see in Act 2 — increasingly to learners themselves, who are positioned as consumers making choices in a market.
That shift in responsibility has consequences. And they fall unevenly.
Act 2: What Does the Apparatus Do to Learners? The Foundation Skills Learner as Consumer.
Peter: The central claim we are making here is that over the last two decades, Australian foundation skills policy has gradually repositioned the learner. Not as a citizen with a right to education. Not as a worker whose development is a shared social investment. But as a consumer — an individual who makes choices in a market, and who bears responsibility for the outcomes of those choices.
That repositioning has consequences. And the most significant consequence is this: when the system is built on the assumption that learners are making free choices, then difficulty, disengagement, or failure can be explicitly and implicitly reframed as personal failing — rather than as evidence of structural barriers the system itself created.
We want to show you how that repositioning happened, and what it looks like in practice. Garry?
Garry: The document where this becomes most explicit is the National Foundation Skills Strategy for Adults, 2012 to 2022. This was Australia’s first national foundation skills policy since 1991 — a significant moment, and one worth examining carefully. Because embedded in it is a particular picture of who the foundation skills learner is, and what they need.
The Strategy makes a set of assumptions about learners that are worth naming directly. It suggests that foundation skills learners:
- are often unaware of the benefits of improving their skills — things like sustainable employment, better wages, improved health, being able to support their children’s learning;
- may be held back by stigma — formed by the traditional image of a “foundation skills class” – associated with primary and secondary schooling
- want to improve, but face day-to-day life pressures that make participation difficult;
- will need more tailored support than a one-size-fits-all approach can provide.
Now — some of this is accurate. These are real challenges.
The Strategy recommended an adult-centred approach supported by a responsive sector, support for learners from families and others who can influence, and support for employers.
This is supported notably by awareness-raising and high-quality learning and outcomes.
But notice what this framing does. It positions the learner as someone who needs to be motivated, informed, and guided toward a market of available options. The system provides the market. The learner’s job is to navigate it wisely.
Providers, in this picture, are responsive businesses — competing for enrolments, tailoring their offerings, meeting learner needs through the discipline of the market. If learners are well-informed and providers are responsive, the theory goes, good outcomes will follow naturally.
It’s a tidy picture.
Let me show you where it breaks down — and I want to do this through the kinds of situations that I think many of you will recognise.
Navigation. The Strategy imagines a learner who can identify their needs, find appropriate providers, and engage with enrolment processes. But those processes — USIs, pre-training reviews, visa requirements, digital systems, complex forms — present significant barriers for people with low foundation skills. The system designed to help them requires the very skills they’re there to develop, just to get through the door.
Support networks. The Strategy assumes learners have access to informed support — family, friends, community members, interpreters — who can help them navigate the market. Sometimes that’s true. Sometimes the market is highly complex for even the most familiar. And in practice, learners may agree to requirements — sign forms, accept enrolments — without fully understanding what they’re agreeing to.
Thin markets. The consumer model assumes choice — multiple providers, flexible delivery, options that can be matched to individual needs. In large metropolitan areas, that might be reality. But in regional and remote areas, choice is often misleading. One provider, significant travel distances, programs that only run if minimum numbers are met. The market doesn’t reach everywhere. And where it doesn’t reach, the model’s promise evaporates.
Viability. This is one that practitioners deeply understand. Even where programs exist on paper, they only run if the numbers stack up. A learner might identify exactly the right course for their needs — and it simply doesn’t eventuate.
Organisational pressures. Training organisations themselves are operating under real financial strain — staffing costs, compliance requirements, facility pressures. These realities shape what providers can actually offer, regardless of what the policy imagines. Flexibility and responsiveness are aspirations that compete with business survival.
Peter: And underneath all of these specific failures is something more structural, which Garry has been pointing toward.
The broader life circumstances that shape whether learning is possible — housing instability, transport barriers, health pressures, financial stress, legal complexity, caring responsibilities — are acknowledged in policy documents. They appear, briefly, as context. But they are then consistently overshadowed by a focus on employment outcomes and market participation.
What this means, in practice, is that the apparatus has limited capacity to respond to the actual conditions of learners’ lives. It can offer a course. It can provide information about choices. What it cannot do easily — within its own logic — is address the reasons why making and sustaining those choices is so difficult.
And here is the consequence we named at the start.
When a system is built on the premise that learners are choice-making consumers, and a learner doesn’t engage, or drops out, or doesn’t achieve — the system’s own logic suggests the explanation lies with the individual. They made poor choices. They weren’t sufficiently motivated. They didn’t take advantage of what was available.
The structural barriers — the thin markets, the navigation demands, the life pressures, the viability constraints — tend to recede from view. What’s left is a story about individual responsibility.
And that is a story with significant consequences and enormous challenges for many of its characters.
Act 3: Where Are We Heading? A History of the Future of Foundation Skills.
Peter: We want to close with something different in register from what has come before.
Acts 1 and 2 were, in essence, a history and a critique. We traced where the apparatus came from, and we showed what it does to learners when its assumptions don’t hold. That work is important — but it’s backward-facing. And the world that Craig’s question pointed toward isn’t behind us. It’s arriving.
So in this final Act we want to do something more speculative. We want to sketch what we called in the book a history of the future — to identify the forces and ruptures that we think a historian of foundation skills, writing in the 2040s, would look back on as the moments that mattered. The turning points that reshaped what the “problem” of foundation skills was understood to be, and who was understood to be responsible for solving it.
We’ll move through three of those moments. And then Garry will bring us back to Craig.
The first moment: the pandemic and the promise of a skills-led recovery.
The COVID-19 pandemic was a rupture in the genuine sense — it forced a rapid, global rethink of work, of who does it, of what skills it requires, and of who gets left behind when the structures that organise work suddenly collapse.
The response from international bodies was, initially, ambitious. The World Economic Forum described the pandemic as an enormous opportunity to embark on an upskilling revolution — a chance to retrain workers displaced by disruption and to build more resilient, capable workforces.
In Australia, the Albanese government’s Jobs and Skills Summit in 2022, and the Employment White Paper of 2023, were the domestic expressions of that ambition. Moments of genuine policy energy, responding to a labour market that had been dramatically reorganised.
But here is what we noticed. And it matters.
The second cycle of the OECD’s PIAAC survey — the international benchmark that had previously done so much to define the “problem” of foundation skills — was due for release in late 2024. It was conspicuously absent from both the Summit and the White Paper discussions.
That absence is not accidental. Who gets to define the problem of foundation skills, and through what measures, and in whose interests — these are always political questions. When a major piece of evidence is available and not used, it’s worth asking why. What picture of the problem did it produce that wasn’t useful to the story being told?
We don’t have a definitive answer. But we think it’s a question that our historian friend might ask in 2040.
The second moment — and the one that we think will be seen, in hindsight, as the most significant — is a specific date.
November 30, 2022. ChatGPT is released into the wild.
We argue in the book that this date represents a genuine turning point in how the problem of foundation skills gets framed — and in what responses become possible, or thinkable, or fundable.
To get a sense of what we were dealing with, we asked ChatGPT itself to define the “tech-bro” — the cultural figure increasingly shaping conversations about AI and its role in education and work. Its response was, we thought, pretty candid:
“A ‘tech-bro’ is often a complex individual characterized by entrepreneurial zeal and technical acumen, coupled with competitive, sometimes ethically questionable behavior… innovative and transformative, yet perpetuating exclusionary practices and overlooking the broader impact of technological advancement.”
A machine intelligence system that can describe its own cultural context with that degree of clarity is worth taking seriously. And worth being cautious about.
What’s emerging from government in this context is a framework of what one recent Australian study called Gen AI workforce personas — categories of worker defined by their relationship to generative AI. At one end: leaders, professionals, and change drivers who govern, build, and deploy AI systems. At the other end — and this is the category that matters most for this room — Gen AI affected workers, whose roles are disrupted by AI implemented elsewhere in their organisations, and Gen AI informed citizens, which includes the unemployed, those outside the labour force, and workers in roles where AI use isn’t expected, but who will increasingly need digital and AI capabilities to participate in work, study, and community life.
Notice where foundation skills learners appear in that framework. Not as leaders. Not as professionals. As the affected and the informed — the people at the receiving end of decisions made by others.
The framework presents this as a challenge for education systems: prepare people to operate effectively in an AI-enabled environment. That’s not wrong. But it is incomplete. And Garry will say more about what it leaves out.
Garry: What it leaves out is the question that a practitioner — or a learner — is entitled to ask.
Not just: how do we prepare people for an AI-enabled world?
But: what kind of literacy are we actually cultivating — or displacing.
I want to bring in a voice here that I think captures something important about the unease that many educators are feeling. Miriam Reynoldson, writing in 2025, opens a lecture on AI and education with deliberate provocation:
‘I’m giving this lecture because I am so, so, so utterly sick of the calls for embedding “AI literacy” into education and workforce development on the grounds that generative AI capabilities are somehow now essential for participation in the world.
First of all, they aren’t essential.
Second, they aren’t literacy.’
Reynoldson’s argument is that the widespread use of generative AI risks eroding the very practices that literacy has historically enabled — reading, writing, making meaning, expressing yourself in the world. That when we outsource the work of communication to algorithm, we aren’t extending literacy. We may be bypassing it. And in bypassing it, we may be foreclosing something that matters — not just instrumentally, in terms of employment outcomes, but in terms of what it means to participate fully in a social world.
That’s a serious argument that deserves to be considered.
And then there is Craig.
Craig, you’ll remember, was excited. He wasn’t anxious about ChatGPT — he was energised by it. For him, the technology offered speed, confidence, and immediate practical feedback. It reduced the risk of mistakes in the written communications that mattered to his business. It gave him a kind of agency he hadn’t previously had access to.
From Craig’s perspective, AI didn’t diminish his literacy. It extended it. It provided scaffolding, modelling, and support that allowed him to participate more effectively in the communicative demands of his working life.
So here we have two positions that are both, in their own terms, coherent.
Reynoldson says: AI literacy isn’t literacy, and the rush to embed it risks displacing the real thing.
Craig says: this tool changed what I can do, and that matters.
We don’t think these positions simply cancel each other out. We think the tension between them is precisely where the important questions live.
For some learners, generative AI will be genuinely enabling. For others, it will represent a further outsourcing of capacities they were never given the chance to develop. The apparatus we described in Acts 1 and 2 — built on competition, commodification, and individual responsibility — is not well-designed to make that distinction. It is designed to scale, to measure, and to credential. It is not designed to ask what kind of person is being formed in the process.
Peter: So we close with the question that we think a historian of the 2040s would be asking about this moment.
Not: did Australia have enough AI-literate workers?
But: when the technology arrived and the frameworks were built and the personas were assigned — what happened to the people at the bottom of those frameworks? What happened to the Craigs?
Did the apparatus that had spent two decades repositioning foundation skills learners as consumers in a market find a way to serve them in a genuinely new environment? Or did it do what apparatuses tend to do — absorb the new disruption into its existing logic, rebrand the product, and leave the structural questions unasked?
We don’t know the answer. We wrote the book in the middle of that uncertainty, and we’re living in it now.
What we do know is that the question Craig asked — maybe one day people don’t need to learn writing because of ChatGPT — was not a naive question. It was a precise one. And it deserves a more careful answer than the one Garry gave him in the moment.
That’s what this book is an attempt to provide. And it’s what we hope this conversation today can continue.