Is a free preprint open access?

It can provide an open access route to a research manuscript, but its review status and reuse terms should be identified separately.

Is open access the same as open data?

No. Publications and underlying datasets are different outputs; access to one does not guarantee access to the other.

Does an open-access article lose copyright?

Not necessarily. Open licenses commonly use copyright to grant permissions under stated conditions.

AI Glossary / Data engineering and architecture / Beginner

Open access

Definition

Open access is the online availability of scholarly work without a reader payment barrier, with reuse permissions an important part of fuller definitions. It aims to make research accessible beyond the institutions or individuals able to pay for subscriptions or individual articles.

What it means

A reader can access the work through a publisher, a repository or another authorized route. Check the version and permissions rather than inferring them from a search-result label.

The Budapest Open Access Initiative describes a broad vision that includes both free access and extensive permitted reuse. In practice, some freely readable material has more limited or unclear reuse terms. It is useful to distinguish the ability to read a work from permission to redistribute or adapt it.

Open access does not mean that a publication is automatically reliable, peer reviewed or current. A preprint, an accepted manuscript and a final published article can have different review and correction histories. Access and evidential quality are separate questions.

How it works

One route is publisher-provided access, often called gold open access. Some publishers fund this through article processing charges, while others use different support. Gold does not by itself mean that every author pays a fee.

Another route is repository deposit, commonly called green open access. An author may deposit a permitted version under applicable conditions; an embargo can delay when the full text becomes available. Diamond open access refers to publishing arrangements with no fees for readers or authors, supported through other resources.

Licenses explain permitted reuse and conditions such as attribution. For example, a CC BY license allows broad reuse subject to its terms, but a site's general accessibility does not establish that every embedded item carries that license.

Stable identifiers, clear version labels and links to corrections help readers locate and evaluate the appropriate work. Making a file temporarily downloadable is not the same as ensuring dependable long-term access.

Three practical examples

A student without a journal subscription reads a research article through an open repository.
A company analyst checks an openly available paper's methods and limitations before using it as evidence in a project review.
A researcher publishes through a diamond open-access journal whose operations are supported without reader or author charges.

Think of it this way

Think of a library making a research collection available to every visitor without charging an entrance fee, while clearly explaining how materials may be copied or adapted.

The analogy is limited because online licenses, versions and third-party content can have different conditions. Free entry answers the access question; it does not settle every question about reuse or the quality of the material.

Worked example

Counting accessible papers without double-counting versions

An invented reading list contains twenty distinct research papers. Twelve are freely available on their publishers' sites. A repository search finds authorized copies of five additional papers and also finds repository versions of three papers already included among the twelve.

The number of distinct papers with an identified free access route is 12 + 5 = 17, or 85 percent of the list. Adding all eight repository results to twelve would incorrectly count three papers twice and suggest complete coverage.

The analyst records the available version and reuse information for each of the seventeen. Finding a repository manuscript does not necessarily mean it is identical to the final publisher version, and finding free text does not establish unrestricted reuse.

The example measures access coverage for a particular list. It is not a claim about the proportion of research that is open access generally, nor a quality score for the papers found.

Illustrative case study

A fictional business-learning publisher prepares an explainer on AI evaluation. Its editor finds a freely available preprint and a later peer-reviewed article describing related results. The team checks the version history and uses the later article's corrected figures where applicable.

Before adapting a diagram, the editor examines its license and credit line rather than assuming the surrounding article's accessibility settles the issue. The team writes an original explanation, cites the evidence and retains links to the sources readers can inspect.

The publisher checks for correction or retraction notices. Open availability makes inspection easier, but does not replace editorial judgment. The case illustrates how access, reuse and evidence quality need distinct checks.

Common misconceptions and mistakes

Open access means no one pays to publish. Publishing still has costs, funded through different arrangements.
Every gold journal charges authors. Some do, while others use different funding models.
Freely readable means public domain. Copyright and license conditions can still apply.
Open access proves peer review or quality. Review status and research merit must be evaluated separately.

Business use

Open access can help organizations consult research without relying exclusively on subscription coverage. It also allows readers to inspect the evidence behind an explanation, report or product claim.

For a knowledge website, record stable source links, publication versions and any relevant reuse terms. Prefer citations that let readers reach the evidence directly. If a repository version differs from the final publication, make that distinction clear where it affects interpretation.

For AI-assisted research workflows, do not equate accessible text with blanket permission for every intended use. Establish the relevant terms for the actual material and workflow. Also evaluate retrieval quality: a system that finds many freely available papers can still overlook relevant evidence or overrepresent particular sources.

Check your understanding

Twelve papers are freely available from publishers. A repository supplies eight results, three of which are versions of those same twelve. How many distinct papers have been found?

Answer: Seventeen. The five additional papers increase coverage; the three overlapping versions should not be counted as new papers.

Related terms

Open science
repository
preprint
accepted manuscript
open licence
gold open access
green open access
diamond open access

Learn next

Read research provenance and open licenses. For three sources you use regularly, identify the version, review status, persistent link and stated reuse conditions rather than relying only on whether the download button works.

Sources

Budapest Open Access Initiative, Declaration: https://www.budapestopenaccessinitiative.org/read/. Foundational primary statement of access and reuse aims.

UNESCO, Diamond Open Access: https://www.unesco.org/en/diamond-open-access. Primary institutional explanation of diamond, gold and green routes.

Creative Commons, CC BY 4.0 deed: https://creativecommons.org/licenses/by/4.0/. Primary licence summary illustrating reuse with conditions, not a statement about every open-access work. The reading-list arithmetic and publisher case are original illustrations.

Previous: Opacity Next: Open class