Política sobre el Uso de Inteligencia Artificial Generativa

General Provision

As a general rule, no article may be written by AI or base the entire scientific work on generative models. Generative AI must be regarded as one tool among the technologies available, but it cannot replace intellectual work, the final review of the submitted document —an essential requirement for being considered an author— or any other activity that is properly authorial in nature. In cases where this technology has been used for tasks that can be automated, a comprehensive declaration must be provided.

Mandatory Declaration and Transparency

The journal permits the use of generative artificial intelligence tools, provided that their use is declared clearly, sufficiently, and verifiably. Transparency in the use of these tools is part of the scientific integrity of the manuscript, as it makes it possible to understand how the text was produced, which technological supports were involved, and how the authors reviewed and incorporated those outputs into the final version of the work.

Any use of generative AI in the preparation of the manuscript must be reported at the time of submission. This declaration must indicate, at minimum, the tool or model used, the version when this information is available, the date or period of consultation, the purpose of use, and a concrete explanation of how the generated outputs were reviewed, corrected, and incorporated into the final manuscript.

This criterion is grounded in the Heredia Declaration, especially in its principle concerning the role of authorship, which emphasizes the need to declare the use of AI and to specify the human intervention applied to generated content. Likewise, following the recommendations of WAME (World Association of Medical Editors), when prompts or input instructions have had a relevant influence on the results, the structure of the text, the analysis, or the interpretation of the work, they must be included or made available to the journal.

The declaration must allow the editorial team, reviewers, and readers to clearly understand which part of the process received AI support and what intellectual work was carried out by the authors.

Exclusion from Authorship and Human Responsibility

Artificial intelligence cannot be considered an author or co-author of a manuscript. Authorship entails intellectual, ethical, and legal responsibility for the published content, and that responsibility can only be assumed by human persons.

This policy is established in accordance with the position statement of the Committee on Publication Ethics (COPE) and with the recommendations of the ICMJE (International Committee of Medical Journal Editors), which indicate that AI tools do not meet the necessary conditions to be accountable for the integrity, accuracy, originality, and scientific validity of an academic work.

Authors are fully responsible for all manuscript content, even when they have used AI tools during its preparation. This includes the review of data, citations, references, arguments, translations, results, tables, figures, and conclusions. Authors are also responsible for verifying that the text does not contain fabricated information, nonexistent references, unsupported claims, unacknowledged biases, or errors produced by the tool.

The COAR Statement on AI and Open Science warns that AI systems may produce erroneous or apparently plausible information without sufficient grounding. For this reason, the use of AI must always be accompanied by responsible human review, academic validation, and ethical control of the content.

Restrictions on Images and Graphics

The journal does not permit the use of artificial intelligence to generate, fabricate, alter, or manipulate images, figures, graphics, or visual materials that have scientific, empirical, or evidentiary value within the manuscript.

This restriction seeks to protect the fidelity of the information presented, intellectual property, data integrity, and trust in the published results. Scientific images, figures, and graphics must faithfully represent the data, materials, observations, procedures, or results of the study.

This criterion is grounded in the Heredia Declaration, especially in its general principle on the protection of intellectual property and the prevention of visual misinformation. It is also supported by the review of editorial policies presented by the Revista Caribeña de Psicología, which notes that international academic publishers, such as Springer Nature and Taylor & Francis, have established restrictions on the creation or manipulation of figures using AI, due to the lack of guarantees regarding their correspondence with the reality of the study.

The journal may request the original files, source data, production records, or documentation needed to verify the traceability of images, graphics, or figures whenever it deems this necessary.

Standardization of the Declaration (AI-Use-12 Taxonomy)

To promote clear, comparable, and traceable declarations, the journal adopts the AI-Use-12 (AI Use Disclosure for Research Articles) taxonomy as a reference framework, as proposed in the specialized literature to organize the reporting of AI use in research articles.

This taxonomy makes it possible to identify with greater precision the ways in which AI may have intervened in the manuscript preparation process, for example, in tasks related to conceptualization, literature search or synthesis, methodological support, data analysis, code generation, visualization, writing, translation, linguistic editing, interpretation of results, discussion, conclusions, or human validation.

Authors must use the AITD Generator – Journal Ready (v1.1) web application to prepare their transparency declaration on the use of AI. This tool makes it possible to organize the information in a structured manner and helps ensure that the report is understandable to the editorial team, reviewers, and readers.

The use of this instrument is linked to the principles of Open Science promoted by SciELO and DOAJ, as it strengthens traceability, editorial transparency, and the responsible communication of the procedures used in the production of scientific knowledge.

The generated declaration must be reviewed by the authors before submission, so that it faithfully reflects the actual use of AI in the manuscript.

Confidentiality and Evaluation

Editors and reviewers must protect the confidentiality of manuscripts, data, supplementary files, review reports, and editorial communications received during the evaluation process.

Following the guidelines of the NIH and the NSF, it is prohibited to upload manuscripts, unpublished data, materials under evaluation, review reports, or any confidential information to AI services not approved by the journal. This measure safeguards the intellectual property of the authors, the confidentiality of the editorial process, and the integrity of peer review.

The use of AI tools by reviewers or editors may take place only when there is express authorization from the journal, sufficient guarantees of confidentiality, and a clearly delimited purpose. Under no circumstances may AI replace the academic, ethical, and editorial judgment of the persons responsible for the evaluation.

Peer review, editorial recommendations, and decisions of acceptance, rejection, or request for revisions must be carried out by human persons. Automated tools may support auxiliary or technical tasks when authorized by the journal, but final responsibility for the editorial process will always rest with the journal’s human editorial team.

Submitting the Declaration of AI Use

Open the web application, complete it as appropriate for your case, and attach the declaration together with the other documents required by the journal: https://ssantamarina.codeberg.page/