The next phase of openness
Open knowledge and data commons were once framed largely as a technical and legal problem. If governments, universities and publicly funded institutions published information in reusable formats under permissive licences, innovation and accountability would follow. That argument was never wholly wrong. Open mapping has improved emergency response and local planning. Open scientific repositories have widened access to research. Public registers and procurement data have given journalists, watchdogs and citizens new ways to inspect power.
Yet the easier part of the project was often the initial release. The harder part is what comes after: keeping datasets updated, interpretable and trusted; deciding what should remain restricted; building institutions that can absorb the costs of stewardship; and ensuring that openness does not become a pretext for extraction by better-resourced actors. The contemporary debate is therefore less about whether openness is desirable in principle than about the terms on which it can remain socially useful.
Open knowledge succeeds when publication is treated as the start of stewardship, not the end of compliance.
This shift matters because data has become a foundational input into administration, research, commerce and civic life. When public information is absent, inaccessible or brittle, decision-making narrows to those with private access. When it is open but poorly governed, value accrues unevenly and trust can erode. The central policy challenge is to build commons that are not only open, but resilient.
From disclosure to public infrastructure
The early open-data movement borrowed heavily from the language of transparency. Publish more information, lower barriers to access and accountability will improve. That logic still has force. The Organisation for Economic Co-operation and Development has repeatedly linked enhanced access to government data with innovation, public-sector efficiency and better service design. The European Union’s framework on open data and the re-use of public-sector information similarly treats high-value datasets as part of the enabling architecture of the digital economy.
But infrastructure is not simply an archive. It requires reliability, standards, financing and maintenance. The lesson from digital public goods is that code, data and documentation degrade without institutional support. A portal launched with fanfare can quickly become a graveyard of static spreadsheets, broken links and undocumented fields. In this sense, many open-data programmes have resembled capital projects without maintenance budgets.
The distinction between disclosure and infrastructure is more than semantic. Disclosure can be episodic and politically opportunistic. Infrastructure must be routine. A transport timetable, land registry, air-quality feed or biodiversity catalogue only becomes broadly useful when users can rely on continuity and consistency. The public value lies less in the symbolic act of publication than in the cumulative effect of dependable access.
Why commons fail in practice
Knowledge commons tend to fail for mundane reasons rather than ideological ones. Metadata is incomplete. Definitions change without versioning. Licensing terms are ambiguous. APIs are introduced, then deprecated. Data is released at a level of aggregation too coarse for serious analysis, or at a level of granularity too fine to manage privacy risks. In other cases, datasets remain formally open but practically unusable because they lack documentation, provenance or machine-readable formats.
The result is a familiar asymmetry. Large institutions with technical capacity can clean, merge and operationalise messy public data. Smaller civic organisations, local researchers and community groups often cannot. Openness then exists in law but not in effect. A commons that can only be productively used by those with substantial resources is not fully public; it is merely non-exclusive in theory.
Open knowledge succeeds when publication is treated as the start of stewardship, not the end of compliance.
This problem has been recognised in research on common-pool resource governance, which has long stressed that shared resources require clear rules, monitoring and collective maintenance. Digital resources differ from fisheries or forests because they are non-rivalrous in use, yet they still depend on scarce labour and institutional attention. Curation, quality assurance and dispute resolution are limited goods. When no one is accountable for them, the commons decays.
Governance is the hidden layer
The most important variable in open knowledge is often invisible to end users: governance. Who decides what is collected, how it is described, under what conditions it is shared and who can contest errors? Which communities are consulted before publication? How are harms assessed? Who pays for long-term preservation? These questions determine whether a data commons serves the public or simply externalises costs.
The most mature approaches increasingly treat governance as a design problem rather than an afterthought. The aim is not to maximise openness at all times, but to calibrate access in line with public value, rights and risk. This may involve tiered access, trusted research environments, independent oversight or data stewardship bodies with explicit fiduciary duties. The Centre for International Governance Innovation and the Ada Lovelace Institute have both argued, in different ways, for moving beyond binary debates of open versus closed towards accountable models of stewardship.
The decisive issue is not whether data is open or closed, but who governs access, under what rules and for whose benefit.
Such an approach is especially important where the subjects of data have historically had little influence over how information about them is used. Indigenous data governance offers one of the clearest examples. The principles developed around Indigenous data sovereignty do not reject sharing outright; they insist that communities should shape the conditions under which knowledge is collected, interpreted and reused. For open-knowledge advocates, that is a useful corrective. Not all barriers to access are obstacles to progress. Some are protections against dispossession.
The politics of standardisation
Standards are often presented as neutral tools for interoperability. In practice, they are political settlements encoded into technical form. A schema determines which categories count, which relationships are visible and which ambiguities are flattened. For land, health, education, labour or environmental data, these choices can shape policy itself.
This is not an argument against standardisation. Without shared formats and identifiers, data commons remain fragmented and expensive to use. But standards should be treated as objects of public reasoning. They need consultation, revision and transparent ownership. Otherwise they risk hardening administrative assumptions that are contested on the ground.
The history of statistical systems offers a cautionary lesson. Categories that appear settled are often products of negotiation, simplification and institutional convenience. As data travels further and is recombined more aggressively, the consequences of those choices multiply. A healthy commons therefore requires not only open datasets but open standards processes, good documentation and pathways for challenge.
Privacy is not the enemy of openness
One of the more damaging habits in public debate is to cast privacy and openness as opposites. They are better understood as co-requirements of legitimate information governance. If people believe that data release will expose them to surveillance, discrimination or re-identification, trust in public institutions will weaken and the political basis for openness will shrink.
The remedy is not blanket secrecy. It is disciplined assessment of purpose, proportionality and risk. Statistical agencies and archives have decades of experience balancing access with confidentiality through de-identification, secure access settings, disclosure controls and legal safeguards. Contemporary debates about linked administrative data, mobility data and health records should draw more heavily on that institutional memory.
The decisive issue is not whether data is open or closed, but who governs access, under what rules and for whose benefit.
The World Bank and the OECD have both emphasised that data policy must integrate safeguards from the outset rather than append them after controversy erupts. In practice, that means better risk assessment before release, clearer redress mechanisms when harms occur and stronger communication about how decisions are made. Openness that ignores social licence is fragile. Privacy regimes that forbid legitimate public-interest uses are brittle in a different way. Durable commons require both openness and restraint.
The maintenance economy of public data
Open knowledge is often spoken of as though its principal costs were initial digitisation and publication. In fact, its dominant costs are recurrent. Servers, archivists, standards specialists, domain experts, legal review, user support, version control and preservation all require funding. The economics are awkward because the benefits are diffuse while the maintenance burden is concentrated.
This creates a predictable temptation: treat open data as a one-off deliverable rather than a continuing service. Many programmes have done precisely that. They have celebrated launch metrics while underinvesting in the less glamorous labour that keeps resources current and intelligible. The pattern is visible across sectors, from local-government portals to research repositories.
If open knowledge is to function as public infrastructure, budgeting must reflect lifecycle costs. That means dedicated lines for stewardship, not just publication mandates. It also means recognising and rewarding maintenance work inside institutions. Academic and administrative systems are still better at valuing novelty than curation, despite the fact that the social returns from reliable maintenance are often higher.
A data commons without a maintenance model is not a commons at all; it is a temporary release with uncertain afterlife.
There is a broader lesson here for public policy. States have become accustomed to financing visible digital transformation projects while neglecting institutional capabilities that are less visible but more decisive. Archivists, registrars, librarians and statistical stewards are not peripheral to digital governance. They are among its core professions.
Research, replication and the scientific commons
Nowhere is the promise and fragility of open knowledge clearer than in research. Open-access publishing, preprint servers and data repositories have widened the circulation of scholarship. During the pandemic, rapid sharing of sequences, methods and findings demonstrated the benefits of scientific openness under pressure. At the same time, concerns about reproducibility, poor documentation and inconsistent data management exposed the limits of simply making more material available.
The scientific commons works best when openness is accompanied by norms of quality, citability and reuse. Data management plans, repository standards, persistent identifiers and reproducible workflows are not bureaucratic ornament; they are the conditions that turn disclosure into cumulative knowledge. UNESCO’s Recommendation on Open Science reflects this broader understanding by combining access with collaboration, equity and long-term stewardship.
There is also a geopolitical dimension. If open science is to remain more than a slogan, lower-income institutions need the infrastructure and bargaining power to participate on fair terms. Otherwise open repositories may widen visibility while leaving underlying inequalities intact. Shared knowledge can still be hierarchically organised. A robust commons therefore requires attention to capacity, reciprocity and the governance of downstream use.
Community knowledge and the question of legitimacy
A data commons without a maintenance model is not a commons at all; it is a temporary release with uncertain afterlife.
Much of the most valuable public knowledge does not originate in central bureaucracies. It is generated by local authorities, civil-society groups, citizen scientists, mutual-aid networks and community archives. These sources can fill gaps in official records, particularly in areas such as housing conditions, pollution, biodiversity and local history. Yet community-generated knowledge often sits uneasily within formal open-data frameworks.
The challenge is legitimacy in both directions. Public institutions may question the reliability or representativeness of local data. Communities may reasonably distrust institutions that absorb their knowledge without attribution, consent or benefit-sharing. A mature data commons must bridge this divide by improving provenance, participatory governance and clear terms of reuse.
Here, libraries, archives and museums have a distinctive role. They have practical experience in cataloguing, preservation and public access, but also in questions of context and interpretation. As institutions with civic mandates, they can help mediate between formal records and vernacular knowledge. In an era of synthetic media and contested truth, this curatorial function becomes more important, not less.
What states should build now
If the next phase of open knowledge is institutional, what should governments and public bodies prioritise? First, they should identify a limited set of genuinely high-value datasets and treat them as critical public infrastructure. That means stable funding, clear ownership, open standards, preservation plans and service-level expectations for quality and update frequency.
Secondly, they should establish governance models that are proportionate to risk. Some datasets can and should be fully open. Others warrant mediated access, secure research environments or community consent mechanisms. The objective is neither maximal release nor reflexive restriction, but a defensible settlement between utility, rights and trust.
Thirdly, they should invest in institutional capability: data stewards, archivists, standards experts, records managers and public-interest technologists. Governance cannot be subcontracted entirely to software. It depends on professions, mandates and accountability structures that endure beyond electoral and funding cycles.
Finally, governments should measure success differently. Counting the number of datasets published is easy but shallow. More meaningful indicators would include usability, update reliability, downstream public-interest use, preservation quality, accessibility for smaller organisations and the presence of mechanisms to contest errors or harmful uses. In short, policy should value whether the commons works, not simply whether it exists.
A more realistic politics of openness
The rhetoric surrounding open knowledge has sometimes oscillated between utopianism and retrenchment. In one register, data liberation would by itself produce innovation, accountability and civic renewal. In the other, mounting concerns about privacy, security and misuse invite a return to gatekeeping. Neither position is adequate.
A more realistic politics begins with institutional humility. Information does not govern itself. Data becomes socially useful through standards, interpretation, maintenance and dispute resolution. These are political and administrative functions. They cannot be wished away by licences, portals or APIs.
Yet realism should not slide into pessimism. The case for open knowledge remains strong precisely because concentrated informational power is itself a democratic risk. Public institutions need scrutiny. Research needs reproducibility. Communities need access to the evidence that shapes their lives. Markets function better when core information is not artificially scarce. The task is therefore not to abandon openness, but to grow up about it.
That means treating the data commons less as a symbolic commitment and more as a constitutional one: a durable arrangement for how societies create, preserve and share knowledge in the public interest. Where such arrangements are absent, information will still circulate, but under terms set by the already powerful. Where they are present, openness can become more than an aspiration. It can become a governing capacity.
