A Learning Commons Beyond Institutional Ownership

Education needs institutions. But it should not belong to them.

Why this vision is needed

I have worked across vocational, community and higher education. Across those settings, I have repeatedly seen thoughtful people create valuable learning opportunities. I have also seen the purposes of learning become entangled with the needs of the organisations that provide it.

Institutions need revenue, enrolments, credentials, grants, regulatory approval and evidence that they are compliant. None of these needs is inherently illegitimate. The problem arises when sustaining the provider is treated as equivalent to serving learners and society.

Compliance can then become a virtue in its own right. What is measurable is treated as accountable; what is accountable is treated as responsible; and what is responsible is assumed to be educationally worthwhile. In Australia, this concern runs across both higher and vocational education: TEQSA regulates higher education, while ASQA is the national regulator for most vocational education and training providers (ASQA, n.d.). The result can be administrative burden that diverts time and resources from teaching and research, encourages defensive documentation and reduces room for patient, relational or exploratory work (Woelert et al., 2026; TEQSA, 2025).

Some of the most important educational activity is difficult to capture in conventional measures: helping somebody recognise a possibility, reconsider a direction, connect prior experience with new understanding, or regain confidence as a learner. It may not produce an immediate enrolment, qualification or revenue stream. That does not make it less valuable (Dymock & Billett, 2008; Foley & Ollis, 2024).

As I approach retirement from academic work, after a relatively short academic career following work in industry, business and government, I find myself increasingly interested in a different organising question:

What might education look like if learning, rather than the institution, were treated as the enduring centre?

What: a distributed learning commons

A learning commons would be a connected ecology rather than another provider. It would not abolish institutions, credentials or standards. It would change what they are for.

Learners, educators, communities, libraries, workplaces, professional groups, public agencies, existing institutions and technological systems could all contribute. No single organisation would own the learner’s journey or need to control every part of it.

A person might learn with a community educator, use resources from a public library or open repository, develop capability through a workplace or community project, receive assessment from an independent source, and retain the resulting evidence in a portable learning record. Later, that learning might contribute to a qualification—or it might remain valuable without being converted into one.

Existing universities, vocational providers and community organisations could remain important nodes. The difference is that the node would no longer be mistaken for the system.

There are partial precedents. Research on learning cities describes locally driven partnerships among universities, educators, citizens and community organisations, while UNESCO frames learning ecosystems as connections across formal, non-formal, informal and experiential learning (Mahmud et al., 2026; UNESCO Institute for Lifelong Learning). Melton’s Community Learning Board offers an Australian example linking community participation, local government, schools, universities, vocational education and other services. Its advisory status also shows the gap between a learning ecosystem and a genuinely learner-governed commons (UNESCO Institute for Lifelong Learning, 2021).

Learner-held continuity

The learner would retain meaningful control over a portable record of learning. Depending on context, it might contain:

• formal qualifications and assessed modules;
• workplace and vocational capability;
• community contribution;
• projects and artefacts;
• recognition of prior and informal learning;
• reflective accounts;
• evaluations from multiple credible sources.

This would not reduce a person to a collection of microcredentials. The record should help preserve continuity, context and interpretation—not merely accumulate badges.

Continuity means that a learner can carry a coherent account of learning across providers, workplaces, communities and life stages. A qualification may be part of that account, but so might a project that changed a person’s direction, a workplace capability developed outside formal study, a community contribution, an educator’s assessment, a self-authored reflection, or evidence that a learner chose not to pursue a credential because another pathway mattered more.

Ownership does not mean that learners can simply invent claims about themselves. Public trust still requires evidence, credible assessment and the ability to verify where a claim came from. But learner ownership does mean that the record is not trapped inside one provider’s database, employment system, platform subscription or credential marketplace. The learner should be able to export it, explain it, withhold parts of it, share different views for different purposes, attach context, correct errors, and move without losing the story of prior learning.

Comprehensive Learner Record pilots and interoperability standards demonstrate that richer digital records can combine assessed learning from curricular, co-curricular, workplace and other settings, including evidence from multiple issuers (Baker & Jankowski, 2020; 1EdTech, 2025). But portability is not the same as learner ownership. A record can be technically portable while still enabling surveillance, context collapse, employer overreach, algorithmic sorting or platform lock-in. Meaningful control requires durable access, exportability, consent, selective disclosure, contextual evidence, correction rights, privacy protections and governance that prevents the record from becoming a permanent dossier.

Learning relationships as infrastructure

Public support would not flow only through institutional enrolments. Some funding could support:

• learner time and access;
• sustained educator–learner relationships;
• independent educational guidance;
• recognition of prior learning;
• open learning resources;
• community-defined projects;
• learners whose needs are unlikely to create institutional profit.

Money would enable learning. It would not be treated as proof that learning had value.

Individual learning accounts provide one partial precedent. Comparative OECD work shows that public training rights can be held by individuals and carried across changes in employment, including by people who may be excluded from employer-funded learning. But the same evidence shows that money and formal choice do not by themselves produce equitable participation. Simple access, independent guidance, targeted support, trustworthy provider information and continuing evaluation are part of the mechanism, not optional additions (OECD, 2019; OECD, 2025).

Singapore’s SkillsFuture Credit demonstrates that a personal learning entitlement can operate nationally, although its participation figures cannot establish that the credit alone caused increased training (Singapore Ministry of Education, 2024). The failure of the United Kingdom’s first Individual Learning Accounts scheme is equally instructive: weak planning, course assurance and system security enabled serious fraud and provider abuse (National Audit Office, 2002). Learner direction therefore requires public safeguards, not simply a marketplace of approved courses.

Warranted trust rather than compliance as virtue

A commons still requires accountability. Learners must be protected from fraud, unsafe practice, exploitation, discrimination and worthless claims.

The alternative is not “no regulation”. It is regulation that is proportional to risk and demonstrably connected to educational or public benefit.

Australian higher-education law already provides a partial precedent. TEQSA must observe principles of regulatory necessity, reflecting risk and proportionate regulation: its actions should not burden a provider more than is reasonably necessary and should be proportionate to non-compliance or the risk of future non-compliance (Tertiary Education Quality and Standards Agency Act 2011, ss 13–16). This establishes that strong baseline protection and differentiated scrutiny can coexist. It does not prove that every risk-based system is educationally beneficial or proportionate in practice.

High-risk activity should receive strong scrutiny. Good practice should earn lower routine burden, not permanent exemption. Actual learner experience and work should matter more than the exhaustive production of protective documentation. New requirements should identify what they replace, and compliance processes should be reviewed when their costs grow faster than their benefits.

TEQSA’s development from early “hyper-regulation” towards a more dynamic approach shows that risk-based assurance itself requires feedback and course correction (Blacklock, Baird & Stensaker, 2025). Risk categories can otherwise become rigid, data-hungry and blind to harms they were not designed to recognise (Baldwin & Black, 2010). Risk judgements, documentary requirements and low-scrutiny status should therefore be transparent, contestable and periodically reviewed.

Contribution rather than organisational activity

Evaluation should move beyond asking only what a provider delivered.

It should ask:

• What became possible for learners that was not possible before?
• Did participants gain capability, agency or greater freedom of choice?
• Who benefited, including people who did not complete a formal program?
• What knowledge or capacity remained after the funded activity ended?
• Did learning circulate into families, workplaces and communities?
• Were participants able to question and redirect the process?

Not all of this can be reduced to a dashboard. Numbers remain useful, but so do artefacts, longitudinal accounts, participant testimony, case analysis and public deliberation.

This is not a rejection of measurement. It is a rejection of measuring only what organisations can easily count. A capability approach asks what learners are actually able to do, choose and become, and treats education as both intrinsically and instrumentally valuable (Walker, 2008). Story-based and deliberative methods such as Most Significant Change can surface unexpected outcomes and make value judgements inspectable, provided they are balanced by quantitative evidence, verification and attention to dissenting or absent voices (Dart & Davies, 2003).

There are practical precedents for this kind of evaluation. England’s community learning survey followed learners up to two years after course completion and examined wellbeing, social and community outcomes, work, parenting and attitudes to learning, alongside learner journeys (Department for Business, Innovation and Skills, 2014). A later community-learning mental-health evaluation combined surveys before, during and after learning with depth interviews 6 to 9 months later across 52 providers (Department for Education, 2018). Qualitative life-history research also shows why this matters: community-based adult learning can help some learners develop or recover social capital in ways that are inseparable from life circumstances and relationships (McIntyre, 2012).

For a learning commons, the right evaluation question is therefore not simply “did the activity work?” but “what worked, for whom, under what conditions, through which relationships and supports?” Realist evaluation is useful here because it attends to context, mechanisms and outcomes rather than treating learning as a standardised treatment (Pawson & Tilley, 1997). Used well, evaluation would combine indicators, testimony, artefacts, case analysis, longitudinal follow-up and public deliberation. It would ask whether agency, relational learning and community benefit actually emerged over time, not merely whether a provider delivered what it promised.

Principles and mechanisms

Some parts of this proposal are principles. They should hold across any version of a learning commons: learning should be organised around learner agency and public value; learners should retain continuity across settings; authority should be distributed, contestable and answerable; funding and assurance should protect learning rather than reward provider volume alone; knowledge produced with public support should circulate where possible; and evaluation should attend to capability, relationships and community benefit over time.

Other parts are mechanisms. Independent learning advisers, portable records, selected separation of teaching and assessment, community funding assemblies, open reporting, expiry dates for compliance requirements and AI-supported navigation are not sacred forms. They are testable ways of enacting the principles. If a mechanism fails, is captured, excludes people, creates new bureaucracy or shifts power away from learners, it should be revised or abandoned without surrendering the larger commitment.

Three risks

The proposal is vulnerable in at least three ways.

First, it could be captured. Providers might relabel ordinary recruitment, compliance or marketing as participation in a commons. Technology firms might make learner records portable in name but dependent on proprietary infrastructure. Confident community actors might dominate assemblies or advisory processes. The answer is not to imagine purer actors, but to distribute authority, publish reasons, limit terms, preserve exit rights and make decisions reviewable by people who were not involved in making them.

Second, it could bureaucratise the very relationships it is meant to protect. Portable records could become another reporting burden; quality assurance could demand exhaustive evidence of every interaction; community deliberation could become a ritual that absorbs time without shifting power. The safeguard is to treat every mechanism as provisional. New reporting duties should say what they replace, record only what serves a clear learner or public purpose, and expire unless there is evidence that they improve learning, protection or equity.

Third, it could exclude people while speaking in the language of openness. Learners with more confidence, time, digital access, prior education or social capital may use the commons first and best. Those with caring responsibilities, insecure work, disability, language barriers, trauma, low trust in institutions or limited digital access may be left with nominal rights they cannot exercise. A learning commons would therefore need active inclusion from the beginning: funded time, accessible guidance, multiple entry points, trusted local intermediaries, non-digital options, culturally safe practice and evaluation that asks who is absent as well as who benefits.

These risks do not defeat the proposal. They define the test. A learning commons is only worth the name if it reduces institutional capture without becoming a new bureaucracy, and widens learner agency without quietly favouring those already best placed to navigate complexity.

What stays institutional, portable or distributed

The argument is not that institutions should disappear. Some functions are legitimately institutional: employing staff, maintaining specialist facilities, protecting students on site, awarding qualifications where the institution accepts public responsibility for the standard, supporting research and scholarship, meeting legal obligations, and sustaining communities of disciplinary, vocational or professional expertise.

Other functions should become portable because they belong to the learner’s continuing life rather than to one provider’s system. These include learning records, evidence of capability, recognition of prior and informal learning, reflective accounts, learner goals, independent advice, complaints history where disclosure is appropriate, and the ability to move without losing identity, context or future options.

A third group should be distributed across the commons. No single provider should own the whole of educational guidance, assessment, quality assurance, public learning resources, funding decisions, complaints, evaluation or community priority-setting. These functions require overlapping authority because they are where institutional self-interest can most easily be mistaken for educational value. The practical question is therefore not “institution or no institution?” It is: which responsibilities should an institution hold, which rights should travel with learners, and which forms of authority should be shared so that no one actor can quietly become the system?

How it could work

A learning commons should not depend on finding permanently virtuous leaders. Any durable arrangement must anticipate capture, corruption, self-interest, risk aversion and organisational self-preservation.

Its design would therefore need several protections.

Distributed authority

No single body should control funding, teaching, assessment, records, quality assurance and complaints. Authority should be distributed across overlapping groups that can inspect and challenge one another.

Governance could combine participation by learners, educators, communities, disciplinary and vocational specialists, and independent public-interest representatives. Election, nomination and random selection could each have a place. Terms should be limited and decisions open to review.

This is a version of polycentric governance: authority is held by multiple decision-making centres rather than by one provider, market or state office (Ostrom, 2010). But distributed authority should not mean dispersed responsibility. Commons governance requires visible rules, affected-party participation, monitoring, conflict resolution and nested oversight (Cox, Arnold & Villamayor-Tomás, 2010). It also still needs the state: to set public-interest defaults, provide trusted information, create negotiation arenas, and act when local or provider interests dominate (Mansbridge, 2014). Practical cases show the importance of mapping who holds authority, resources and information, while research on deliberative mini-publics warns that participation can be captured unless agenda-setting, deliberation and formal decision-making are visible and checked (Meinzen-Dick et al., 2021; Junius, 2025).

Contestable decisions

Neither expertise nor majority support should create unanswerable authority.

Public values and technical judgment must be connected but distinguished. Communities should help determine the ends worth pursuing. People with demonstrated knowledge can help assess the likely means, consequences and uncertainties. Both kinds of judgment should remain visible and open to challenge.

Portable rights

A learner should be able to leave one network or provider without losing their history, evidence, identity or future options. Records should be portable, understandable and under meaningful learner control. Technical verification can establish the source and integrity of a claim, but not its truth or educational value. Privacy therefore requires deliberate safeguards against unnecessary disclosure, aggregation and correlation (W3C, 2025).

Open circulation

Resources and knowledge produced with public support should remain publicly accessible wherever privacy, cultural authority and legitimate restrictions allow. Community participation should not become unpaid extraction feeding a proprietary educational product.

Open educational resources provide one model: learning, teaching and research materials can be released under licences that allow no-cost access, reuse, adaptation and redistribution (UNESCO, 2019). Australian public research policy makes a similar public-funding argument: knowledge generated through public grants should be openly available where possible (ARC, 2026). But openness should be a public-interest discipline, not a licence for extraction. Restrictions may be justified for privacy, human-subject protection, intellectual property, personal information, sacred or secret Indigenous knowledge and other legitimate public or community interests (UNESCO, 2021). Indigenous data and cultural knowledge require particular care: the CARE Principles centre collective benefit, authority to control, responsibility and ethics, while Traditional Knowledge and Biocultural Labels show how communities can express protocols for access, use, attribution and future circulation (Carroll et al., 2020; Local Contexts, n.d.).

Educators as stewards

The educator’s role would be less about representing a provider and more about stewarding learning relationships. This might involve helping learners articulate purposes, connecting people and resources, challenging assumptions, interpreting evidence, supporting reflection, and recognising learning across contexts.

Educators might participate in several networks rather than belonging wholly to one organisation.

This kind of work is boundary work, not simply pastoral support. Communities of practice research treats learning and identity as formed through participation in shared practice (Wenger, 1998), while boundary-crossing research shows how brokers and boundary objects can help people move between different social, institutional and occupational worlds (Akkerman & Bakker, 2011). Vocational education already offers a practical example: VET teachers often sustain their professional identity and currency by moving between school and working life, translating occupational knowledge across settings (Andersson & Köpsén, 2019). But educator agency depends on time, trust, resources, professional history, relationships and institutional conditions—not exhortation alone (Priestley, Biesta & Robinson, 2015). A learning commons should therefore avoid turning “stewardship” into an unpaid moral obligation or a managerial slogan (Watson, 2014).

AI as support, not sovereign

Artificial intelligence could help people connect dispersed experiences, test claims, identify contradictions, understand uncertainty, and consider longer-term consequences. It may also help maintain learner-controlled records and make complex evidence more accessible.

But AI should increase answerability without becoming unanswerable. Its recommendations must remain inspectable and contestable. It should not be permitted to turn prediction into authority, learner data into surveillance, or apparent consensus into an instrument of control.

AI in a learning commons should therefore be treated as a governed support system, not a sovereign decision-maker. UNESCO’s guidance on generative AI in education argues for human-centred use, privacy protection, ethical validation, pedagogical purpose and institutional readiness (UNESCO, 2023). European guidance foregrounds human agency, fairness, transparency, privacy, accountability and justified choice (European Commission, 2022). These principles matter most when AI affects access, assessment, progression, funding or complaints. The EU AI Act treats several education and vocational-training uses as high-risk and requires safeguards such as risk management, documentation, logging, transparency, human oversight, accuracy and robustness, with a right to meaningful explanation for some high-impact decisions (European Union, 2024). Australian education policy is moving in the same direction through principles of teaching and learning, wellbeing, transparency, fairness, accountability, privacy, security and safety (Australian Government Department of Education, 2023). In practice, high-stakes AI should have clear purpose limits, disclosure, human review, appeal pathways, audit records, monitoring and non-AI alternatives.

Beginning without waiting for a new system

A learning commons need not begin with wholesale institutional reform. It could emerge through practical experiments such as:

• independent learning advisers who are accountable to learners rather than providers;
• portable learner-owned records;
• stronger recognition of workplace, community, prior and informal learning;
• public funding available to approved networks and individuals, not only registered institutions;
• selected separation of teaching and independent assessment;
• community assemblies allocating part of public education funding;
• open reporting of spending on teaching, administration, marketing, compliance and executive functions;
• review or expiry dates for documentary obligations;
• longitudinal evaluation of learner agency and community benefit;
• qualifications assembled across several learning settings where this is educationally coherent.

These are not a blueprint. They are ways of testing whether learning can become less dependent on institutional ownership while retaining public responsibility.

Existing initiatives already approximate parts of this. Cities of Learning and LRNG connected schools, libraries, museums, employers, community organisations and digital platforms so young people could access learning across a city and have some of it recognised through digital badges and pathways (Urban Institute; The Sprout Fund; RSA). Unionlearn showed how trusted workplace learning representatives could advise, coach and signpost workers into learning through relationships that were not simply provider recruitment (OECD, 2019; Bacon & Hoque, 2009). Big Picture Learning Australia’s International Big Picture Learning Credential shows how portfolio evidence, internships, educator judgement, learner profiles and non-ATAR university pathways can recognise learning that standard assessments often miss (Big Picture Learning; Big Picture Learning Australia).

Each example also shows what remains unresolved. Cities of Learning can become platform- or badge-centred. Unionlearn depends on union infrastructure, employer cooperation and political support. Big Picture remains mostly school-based and does not solve lifelong portability or distributed public governance. The learning commons proposal is therefore not invented from nothing, but neither is it already achieved. It asks whether these partial strengths can be brought together under stronger learner ownership, active inclusion, distributed authority and public-interest evaluation.

What would weaken this proposal?

The learning commons should not be protected from evidence. The proposal would be weaker if distributed arrangements consistently increased inequality, leaving confident and well-connected learners better served while others faced more complexity. It would be weaker if portable records intensified surveillance, employer screening or algorithmic sorting more than they strengthened learner agency. It would be weaker if unbundling teaching, assessment, guidance and assurance reduced educational quality, weakened care, or made responsibility harder to locate when learners were harmed.

It would also be weaker if the administrative work of maintaining the commons became heavier than the institutional systems it was meant to improve, or if community governance was repeatedly captured by the people with the most time, status, money or procedural confidence. Most importantly, it would be weaker if learners did not experience greater capability, continuity, dignity, choice or public value over time.

That is the standard the proposal should accept. A learning commons is not valuable because it sounds more open than an institution. It is valuable only if it can protect learners, widen agency, sustain serious learning and return more educational value to public and community life than the arrangements it seeks to supplement or replace.

The proposition

Education should not primarily be something institutions deliver to populations.

It should be a socially supported capacity through which people learn, contribute, judge evidence, make consequential choices and participate in shaping their worlds.

Institutions may remain useful nodes in that ecology. The error is allowing the node to become the purpose.

This is not an argument against institutions, teachers or standards. It is an argument against confusing educational means with educational ends. UNESCO’s futures work frames education as a public endeavour and common good grounded in rights, solidarity, shared responsibility and accountability to all (UNESCO, 2021; UNESCO, 2015). Locatelli argues that common-good language foregrounds participation, democratic governance and social justice beyond a narrow technical idea of public good (Locatelli, 2018). Biesta’s work adds the educational warning: when measurement, economic purpose or lifelong-learning policy displaces democratic questions of purpose, education can become an individual duty or delivery problem rather than a collective public practice (Biesta, 2009; Biesta, 2006). Marginson similarly cautions that higher education’s public mission disappears when it is treated only as an aggregation of private benefits (Marginson, 2011). Adult education research extends the same point across the life course: lifelong learning becomes a common good only when it is accessible, available, affordable and socially supported (Boyadjieva & Ilieva-Trichkova, 2018).

---

About this piece

This vision developed through an extended dialogue between Dr Michael J. Henderson and ChatGPT. It draws on Henderson’s experience across vocational, community and higher education, and on his research interests in adult learning choice-making, identity, technology and leadership. It also extends thinking developed through his PhD journey and thesis, Adult Learning Choice Making Today: A Journey of Disparate Transitions (Henderson, 2023).

The ideas, judgments and responsibility for publishing them remain Henderson’s. AI was used as a dialogic partner: to question assumptions, identify tensions, organise the emerging argument and assist with drafting.

• Michael Henderson’s website
• ORCID: 0000-0002-4372-8590
• LinkedIn