Open science needs to fund the work of making research usable

Requiring researchers to share their work is easier than paying for someone else to use it. A funder can ask for an open dataset at the end of a project without making room for the explanation, checking and continuing care that give those files value. The resulting deposit satisfies a condition while leaving the next researcher to reconstruct what it means.

Open science policy needs to treat that work as a funded part of research. The obligation falls especially heavily on institutions with limited technical support: the same researcher may have to organise the data, explain its limitations and answer later questions while beginning another assignment. That is a risk in the design of a requirement, rather than a claim that every institution currently handles it this way.

UNESCO's 2021 Recommendation on Open Science already recognises the need for long-term investment in infrastructure, services and people. It also addresses incentives and the inequalities that can arise from scientific publishing models. The policy foundation is broader than free access to a finished paper. Implementation needs to preserve that breadth when a grant is translated into a budget and a reporting form.

There is a timely opportunity to examine the gap. UNESCO's dataset published in August 2026 underpins dashboards based on the 2025 reporting cycle, drawing on responses from 81 member states. The dashboard itself cautions that the information is reported by states and is not independently verified by UNESCO. These records can support questions about implementation; they do not, by themselves, show what a researcher experiences inside a particular institution.

A useful question for funders is what happens between collecting information and making a deposit. Consider a hypothetical small research team completing interviews in more than one language. Sharing the recordings may be inappropriate because of consent and confidentiality. Preparing a suitable alternative could require explaining the interview procedure, documenting how categories were translated and identifying which material cannot be released. Those tasks need time and judgement.

The cost is not solved by telling the team to upload everything. Nor does responsible openness mean exposing participants or releasing information that the team has no right to distribute. A clear account of restrictions can be a valuable research output. It allows another scholar to understand the boundary of the available evidence and, where appropriate, ask about a legitimate access procedure.

Budgets could therefore include a specific, proportionate allocation for preparing reusable material. Applicants would explain what they expect to share, what supporting documentation is needed and who will do the work. Reviewers could assess whether the allocation fits the project instead of treating it as a general administrative expense to remove during negotiation.

There should be room to revise that plan. A project may encounter an unexpected ethical restriction or produce a type of material different from the one anticipated. Requiring a reasoned update would be more credible than rewarding a team for delivering the original file list regardless of what it learned. The objective is a useful research record, with legitimate limits explained.

Repository support also needs a realistic division of responsibilities. A librarian or data steward can help make a collection understandable, but that professional cannot infer missing decisions that were never documented. Researchers need to record the meaning of variables, changes to methods and relevant exclusions during the project. Paying for a final deposit will not recover information that the team has already lost.

The obligation continues after publication, although it cannot continue indefinitely at one person's expense. A project needs to identify where corrections belong and what happens when the original researcher changes institution. The repository should explain the support it offers and the limits of that support. Users can then distinguish a preserved collection from a continuing service that answers questions about it.

There is a strong objection to adding another requirement to already complicated grants. A detailed sharing plan could become paperwork that large institutions complete easily and smaller ones struggle to produce. The answer is proportionate assessment. A modest qualitative study and a major shared database need different arrangements. Funders can provide practical examples and allow pooled institutional services where these are suitable.

Recognition matters alongside money. If preparation and stewardship take time, the people doing them need a contribution that can be described in appraisal and promotion. Naming a responsible contributor also helps users understand whom to credit. Otherwise the policy can depend on work that remains professionally invisible while the more easily counted publications determine careers.

UNESCO's implementation framework includes investment in skills and changes to incentives. Governments reviewing their progress can ask how those commitments appear in actual funding decisions. An inventory of platforms and policies is useful, but it needs to be followed by questions about staffing, maintenance and the conditions attached to a deposit.

The next grant call offers a concrete place to begin. It can make preparation for reuse eligible for support, ask applicants to explain reasonable restrictions and identify who will care for the record after the project ends. Researchers would then receive a means to meet the obligation being placed on them. Open science will be easier to use when its budgets recognise the labour between a completed study and an intelligible collection.

Swaansh Mahat is an independent writer and English-language educator based in Kathmandu, Nepal. He writes about education, public services and everyday institutional life.