Horizon open science: decide what will be open, protected and costed before submission

A proposal that promises to “share results widely” has not yet designed its open-science route. The expected outputs may include publications, datasets, software, models, protocols, methods, standards, training material, policy tools, prototypes or other knowledge assets. Each can have different ownership, access, protection, consent, repository, licensing and timing conditions.

The European Research Executive Agency's current guidance identifies two mandatory Horizon Europe practices: open access to publications and open access to research data based on the principle “as open as possible, as closed as necessary”. The Commission's wider open-science page links responsible research-output management to the FAIR principles: findable, accessible, interoperable and reusable.

This is not a direction to publish every item immediately. It is a requirement to make deliberate, justified and implementable decisions about access.

REA says the proposal methodology should clearly demonstrate how the project will comply with mandatory open-science practices. It also says applicants should outline their future Data Management Plan and show relevant open-science capacity.

A credible methodology therefore needs more than an aspirational commitment. It should identify the main expected output classes, how they will be managed, who will make access and protection decisions, and which work packages carry the effort. If a project expects to collect sensitive human data, build commercial software, generate a patentable result or combine data across jurisdictions, the access route can affect ethics, intellectual property, exploitation, procurement, infrastructure and partner agreements.

Those connections must be visible before the proposal is submitted. Otherwise, the consortium risks promising openness in one section and promising exclusive commercial control, strict confidentiality or unrestricted data transfer elsewhere.

PRINCEPS recommends one row for each material output or output class, with nine fields:

1. Output — publication, dataset, software, model, protocol or other result.

2. Origin and contributors — which partners, sites, people, background assets or third-party inputs produce it.

3. Owner and decision authority — who owns or controls the result and who can approve release, protection or restriction.

4. FAIR and repository route — the metadata, format, standards and trusted repository expected to make the output findable and reusable where appropriate.

5. Access timing — when deposit or public access is expected and what must happen first.

6. Licence or reuse conditions — the proposed publication, data, software or other reuse terms.

7. Restriction and justification — privacy, confidentiality, security, intellectual property, commercial exploitation or another legitimate constraint.

8. Resources — data stewardship, repository, curation, anonymisation, legal review, protection, software maintenance or other effort and cost.

9. Responsible task and update point — the work package, deliverable, owner and date for reviewing the decision.

This matrix is a PRINCEPS proposal-control method, not an official EU template. It prepares the consortium to write a coherent methodology and later maintain the formal Data Management Plan.

REA's guidance says access to research data may be restricted where openness conflicts with legitimate interests, including commercial exploitation, data protection, privacy, confidentiality, trade secrets, Union competitive interests, security, intellectual property rights or other Grant Agreement obligations. It also says that the reason for restricting access should be explained in the Data Management Plan.

The practical lesson is that “confidential” is not a complete access strategy. The consortium should specify which output or field is restricted, the reason, the decision authority, the duration and what metadata or aggregated result can still be shared. A proportionate restriction can protect legitimate interests while preserving as much findability and reuse as the project can responsibly support.

Open science and exploitation are not automatically in conflict. REA notes that patenting can coexist with open-science obligations, but protection decisions need to occur before public disclosure where novelty must be preserved. The proposal should therefore identify the pre-publication review, responsible partner and maximum decision period. An undefined publication hold can undermine open access; an uncontrolled early disclosure can undermine protection.

The same discipline applies to African partners. A site or institution contributing data, local knowledge, implementation evidence or software should not disappear from the ownership and access model. The consortium should state its contribution, authority, benefit, repository role, access rights and any post-project responsibilities. Open science should support equitable collaboration, not become a route for extracting and relocating knowledge without agreed governance.

Before the proposal is frozen, reconcile the output-access matrix with the methodology, ethics section, intellectual-property strategy, exploitation plan, consortium roles, budget and work packages. Confirm that every important output has an owner, access route, protection decision, justified restriction where necessary, repository path, timing and resource plan.

This analysis is strategic and editorial. It is not legal, data-protection, intellectual-property, cybersecurity, research-ethics or grant-award advice. The live topic, application form, Model Grant Agreement, institutional policies and qualified advice remain controlling. Open access, protection, exploitation and funding outcomes are not guaranteed.

Editorial note: This is PRINCEPS analysis for general information. It does not replace the official work programme, topic conditions, submission system, grant rules or professional advice specific to an application.