This page exists because there is a question any informed reader will ask before any other: Ikusmira is a site whose encyclopaedic articles are written with the help of language models, and whose academic editor has a mode that goes from an empty document to a full draft. What, then, separates Cuadernos de Ikusmira from a content farm with a DOI? The answer is not to hide it; it is to measure it and publish it. Esta página en castellano.
Authorship
Authors are people with a name, an institutional affiliation and an ORCID. A language model is never an author or co-author: it cannot answer for the content, give consent or declare competing interests. All signatories have contributed substantially to the work and approve its submission; those who helped without reaching authorship are named in the acknowledgements.
Use of artificial intelligence
It is permitted, and it is declared. Every article publishes an AI-use statement saying which tool was used and at which stage (literature search, drafting, copy-editing, translation, analysis), or that none was used. The sentence “the authors answer for the content” is not boilerplate: it is the criterion.
Three rules that are checked, not promised:
- No fabricated reference can be published. The editor in which manuscripts are written only cites records that exist in eight academic indexes, and submission is blocked if a reference does not resolve to a DOI, an arXiv identifier, a PMID or an OpenAlex record, if a citation points to a reference missing from the list, or if a reference is never cited. This is the most frequent and most serious failure of model-generated text, and in this journal it is closed by construction.
- The assistant proposes, the person decides. In Luma, everything the model writes arrives as a proposal anchored to the text that the author accepts or discards, and the system records what was done with each one. There is no silent generation.
- Reviewers are people. A model may help a reviewer check a citation or understand a method; it does not write the report or decide the recommendation. Reviewers declare their use just as authors do.
Competing interests
Authors, reviewers and editors declare any financial, institutional or personal relationship that could influence the work or its assessment. The author’s statement is published. An editor does not handle a manuscript whose author they have a conflict with; a reviewer with a conflict declares it and is replaced.
Self-publishing
An article signed by a board member is handled by another editor. An article signed by the coordinating editor is handled by an external guest editor and its review reports are published alongside the article. A journal that mostly publishes its own editor is a known silhouette, and the way not to be one is to be checkable.
Plagiarism and originality
We do not use a commercial similarity-detection service. What we do, stated as is: similarity checks against Ikusmira’s own corpus (more than four thousand indexed articles) and a manual sampling of manuscript sentences in a general search engine. Plagiarism, undeclared self-plagiarism and redundant publication are grounds for rejection or, if discovered later, retraction.
Data, code and reproducibility
Data and code underpinning a result must be available, with a persistent identifier, or the availability statement must explain why not. Fabrication or manipulation of data is misconduct and is treated as such.
Misconduct and complaints
Any suspicion of misconduct (plagiarism, fabrication, ghost or gift authorship, undeclared conflict) may be reported to the board through the contact page. The board examines it, hears the author and decides. Decisions affecting a published article (correction, erratum, retraction) are published with their reasons, following the editorial policy.
What we do not do
We do not charge to publish or to read. We do not accept sponsored content. We do not publish an article for the sake of a DOI: the DOI comes after acceptance, not before. And we do not pretend the journal is more than it is: it is new, its board is small, and its first articles are written by those who founded it and those they invite. That is also said on every page.