I. Preamble
The journal acknowledges that artificial intelligence systems can legitimately support certain research and writing activities. However, their use must preserve human authorship, transparency, reliability, confidentiality, data protection, third‑party rights, and the integrity of research and scientific publishing.
This Policy adopts as the national reference CNPq Ordinance No. 2,664, dated March 6, 2026, and incorporates its guidelines as editorial conditions applicable to all submissions received by the journal, regardless of the source of funding.
The CNPq guidelines require the declaration of the use of generative artificial intelligence at any stage of the research, including conception, writing, data analysis, and submission; prohibit presenting generated content as if it were of human authorship; assign full responsibility to the authors; and recommend preserving data, metadata, protocols, code, software, and other research materials.
The journal also adheres to Brazilian copyright legislation; the General Data Protection Law (LGPD); the ethical standards applicable to research involving human subjects; the guidelines of the Committee on Publication Ethics (COPE); the recommendations of the International Committee of Medical Journal Editors (ICMJE); the recommendations of the European Association of Science Editors (EASE); and other scientific integrity standards applicable to the manuscript’s field.
II. The Six Fundamental Rules
When submitting a paper to the journal, the authors must observe six fundamental rules:
- Artificial intelligence tools cannot be authors or co‑authors.
- Any conscious use of generative artificial intelligence must be declared, including in the conception, drafting, translation, data analysis, code development, image production, manuscript submission, or revision.
- The authors must directly verify all sources, references, citations, data, calculations, code, and results.
- Personal, sensitive, confidential, or protected data may not be entered into external tools without proper authorization.
- Synthetic or AI-altered content must be identified and must never be presented as real data, documents, or events.
- The authors remain fully responsible for the work, even if part of it was produced or modified with the aid of AI.
All authors must be aware of and approve the declared AI uses at the time of submission.
III. On the Use of Artificial Intelligence
For this Policy, “use of artificial intelligence” is considered any human action that uses a system capable of producing, transforming, summarizing, translating, classifying, or analyzing texts; references; data; code; images; graphics; sounds; videos; models; recommendations; other content.
The use of artificial intelligence encompasses both autonomous tools and generative functionalities embedded in word processors; browsers; search engines; bibliography managers; translation systems; programming platforms; data analysis applications; image editors; other software.
The obligation to declare is determined by the function actually used, not merely by the product’s commercial name. CNPq expressly includes conception, writing, data analysis, and submission among the phases subject to declaration.
They do not need to be declared, except when they use generative artificial‑intelligence functionality: conventional spell checkers; automatic style application; deterministic formatting; word counting; bibliographic managers used only to organize already verified references; conventional calculators and statistical software; search tools that merely retrieve results, without generating or transforming content.
IV. Full liability of the authors
Only natural persons who meet the journal's authorship criteria may be identified as authors or co‑authors. AI tools may not be included in the list of authors; receive an ORCID identifier; have institutional affiliation; receive an author role in the CRediT taxonomy; assume responsibility for content; approve the final version; be accountable for conflicts of interest; respond to inquiries after publication.
The authors are fully responsible for originality; accuracy; integrity; argumentative coherence; interpretation of sources; data and results; calculations; code; references; citations; images; translations; copyright; data protection; conclusions; statements submitted to the journal.
The use of artificial intelligence does not reduce or transfer the authors' responsibility. Claiming that a particular error, plagiarism, nonexistent reference, distortion, or incorrect data was produced by the tool does not absolve the authors of responsibility.
The authors must be able to explain and defend all the arguments presented; the methodological procedures; the analytical choices; the functioning of the codes used; the meaning of the results; the conclusions of the work; and the references used.
V. From the AI Use Declaration Register
The declaration will occur in two instances, namely: in the submission system and in the manuscript submitted for journal review by the corresponding author.
The manuscript must include, immediately before the references, a section: Statement of Artificial Intelligence Use. This section should list the uses that have influenced the research; the method; the data; the code; the analysis; the interpretation; the writing; the translation; the images; the final content of the work.
When AI is integrated into the method, the essential information must also be presented in the methodological section. If AI is used in a figure, image, audio, video, table, or graph, the corresponding caption or description must also indicate this use.
The statement must be updated whenever there is new use of AI after the initial submission. This includes uses that occur during the revision requested by the editorial team; in drafting responses to reviewers; in new data analysis; in the production or modification of figures; in the translation or rewriting of parts of the text; and in the preparation of the final version. The statement must correspond to the version actually submitted and published.
VI. Of Prohibited Uses
Listing an AI system as author or co‑author is prohibited; concealing its use is subject to disclosure; presenting AI‑generated content as if it were solely human contribution is prohibited; delegating the central intellectual contribution of the work to AI is prohibited; fabricating or falsifying data is prohibited; inventing results, references, citations, authors, works, or documents is prohibited; creating nonexistent legislation, case law, judicial decisions, or administrative acts is prohibited; inventing participants, interviews, statements, testimonies, or field observations is prohibited; presenting synthetic data as if it were real data is prohibited; using paraphrasing or automated rewriting to hide plagiarism or self‑plagiarism is prohibited; citing a work that the authors did not consult is prohibited; attributing content to a source that does not contain it is prohibited; manipulating images or graphics to favor a hypothesis is prohibited; altering visual data in a way that changes its scientific meaning is prohibited; inserting personal, sensitive, confidential, or protected data into an unauthorized tool is prohibited; uploading non‑public manuscripts, opinions, or documents of third parties into AI systems is prohibited; using protected material without proper permission or license is prohibited; claiming that code or analysis has been verified when the authors cannot explain or reproduce it is prohibited; creating deceptive identities, documents, or communications is prohibited; inserting hidden instructions intended to manipulate reviewers, editors, or automated systems is prohibited; using AI to circumvent editorial integrity mechanisms is prohibited; omitting relevant limitations of the tool or method is prohibited.
VII. Sources Used
The authors must directly consult the sources used. Every reference suggested by AI must be verified for the existence of the work; authorship; title; date; publisher or journal; volume and issue; pages; DOI or other identifier; and the correspondence between the work and the presented claim.
Every direct quotation must be compared with the original source. Summaries, paraphrases, and translations must faithfully preserve the meaning of the source.
AI outputs should not be used as a primary source to support scientific, historical, philosophical, legal, or factual claims. EASE recommends verifying the original sources and warns against using AI outputs as the primary basis for statements.
When the system's own output constitutes the object of the research, it may be reproduced or analyzed as empirical material, provided that this condition is explicitly stated; the tool is identified; the date of the interaction is recorded; the relevant prompts and parameters are preserved; and the reproducibility limitations are disclosed.
VIII. Of Data, Code, and Reproducibility
When AI is used on data or code, authors must disclose: tool and provider; model and version, when available; date or period of use; relevant input data; purpose; methodologically relevant parameters or instructions; validation procedures; limitations; bias risks; reproducibility measures.
The authors must preserve, to the extent necessary and permitted, original data; processed data; code; versions; execution logs; relevant prompts; relevant outputs; parameters; documentation.
EASE recommends that uses of AI in data collection, cleaning, analysis, interpretation, or generation of data or code be declared, accompanied by reliability indicators and reproducibility measures. Synthetic data should be clearly identified as synthetic. The journal may request the materials necessary to verify the work.
IX. From Images, Figures, Audio, and Video
Content generated or altered by AI must be clearly identified. It is prohibited to present synthetic content as if it were an authentic record of a person; document; event; experiment; specimen; location; phenomenon; research data.
When an image, recording, or signal constitutes scientific data, AI may not be used to alter its informational content in a deceptive manner.
Operations such as generative filling, image expansion, part reconstruction, removal or insertion of objects, background replacement, cloning, creative restoration, voice synthesis, and facial reconstruction must be declared when used.
The original files must be preserved.
Illustrative figures may be admitted when they are identified as generated or altered by AI; are not presented as evidence; do not mislead the reader; and respect copyright, image, privacy, and dignity.
X. Human subjects research
The use of AI in research involving human subjects must be compatible with the approved research protocol; the consent or other applicable ethical and legal basis; the information provided to participants; the rules of the responsible institution; data protection legislation; and confidentiality safeguards.
Authorization to collect data does not automatically imply permission to transmit it or to allow an external AI provider to process it.
When the use of AI has not been contemplated in the protocol, researchers must assess, before processing, the need for consultation with the competent ethics body; an amendment to the protocol; new authorization; and additional information for participants.
It is prohibited to use AI to fabricate statements, participants, testimonies, responses, interviews, field observations, or nonexistent empirical results.
Legitimate simulations must be explicitly identified as simulations.
XI. Personal, Sensitive, and Confidential Data
Authors must not insert, transmit, upload, or allow access via unauthorized tools to identified or identifiable personal data; sensitive data; data concerning children and adolescents; health information; genetic or biometric data; identifiable interviews; medical records; confidential documents; embargoed materials; professional, commercial, or industrial secrets; third‑party manuscripts or opinions; content subject to confidentiality agreements; databases whose license prohibits such processing.
The LGPD considers processing, among other operations, the access, use, transmission, processing, storage, communication, and transfer of personal data. Even in academic activities, the applicable requirements of articles 7 and 11 must be observed.
The fact that a tool is paid, institutional, corporate, advertised as private, and configured not to train the model does not, by itself, prove that the processing is authorized.
They should be considered, based on the risk: retention, reuse, training, storage, sub‑processors, international transfer, security, access control, disposal, and rights over inputs and outputs.
In case of doubt, the data should not be processed before guidance from the competent institution.
XII. Copyright and Third-Party Rights
The authors are responsible for ensuring that the use of AI respects moral and patrimonial rights; licenses; terms of use; image and voice rights; personality rights; confidentiality obligations; rights over databases; software licenses; and rights over third‑party materials.
AI outputs may reproduce protected material even when they do not disclose its source. Authors must identify the original sources; obtain permissions when necessary; verify whether the content can be published; verify whether the material can be submitted under the license adopted by the journal; disclose any relevant restrictions.
A tool's claim that its output can be used does not prove the absence of third‑party rights.
XIII. Of the Declaration Models
The positive statement must disclose, when applicable, the supplier; tool name; model and version; date or period of use; research phase; purpose; affected parts of the work; nature and extent of the intervention; verification or validation procedures; use of personal, sensitive, or confidential data; environment used; reproducibility measures.
As a general rule, the full publication of the entire prompt history will not be required, but the editorial team may request the prompt history for evaluation of compliance with policy and editorial standards.
Authors must preserve materially relevant prompts and outputs; include in the article the elements necessary for understanding the method; make them available to the journal upon a justified request; protect personal data, confidential content, and third‑party rights.
13.1. Standard AI Use Statement
Artificial Intelligence Use Statement
During the [phase] stage, the authors used [tool, vendor, model, or version] during the period of [date or timeframe] for [purpose]. The usage applied to [sections, data, code, images, or other elements]. The outputs were checked and validated through [verification procedures]. No personal, sensitive, or confidential data were entered into unauthorized tools [or describe the authorized processing]. All authors reviewed and approved the final content and assume full responsibility for its originality, accuracy, completeness, references, data, methods, results, and conclusions.
13.2. AI declaration model as method or object
Artificial Intelligence Use Statement
This study used [tool, provider, model, and version], accessed on [date], as [research method or object]. [Input data, parameters, prompts, and relevant procedures] were employed. The outputs were evaluated through [procedures], and the system’s limitations were considered in [description]. The records necessary for understanding or reproducing the work are available in [repository or supplementary material], subject to the ethical, legal, and confidentiality restrictions described in the article. All authors reviewed and approved the final content and assume full responsibility for its originality, accuracy, integrity, references, data, methods, results, and conclusions.
13.3. Negative declaration
Artificial Intelligence Use Statement
The authors declare that they did not consciously use generative artificial intelligence tools in the design of the research, selection or analysis of sources, production or interpretation of data, code development, generation or alteration of images, writing, substantive translation, review, or submission of this work. Only conventional non‑generative resources were used, where applicable.
The journal may keep this statement only in the editorial record, without necessarily publishing it in the article.
XIV. From the Editorial Evaluation
The journal may examine the coherence of the statement; references; citations; data; code; images; manuscript versions; other elements relevant to integrity.
Detection tools for potentially AI-generated content may be used only as auxiliary indicators.
No score, percentage, or automatic classification shall be considered, in isolation, as proof of AI use; proof of omission; sufficient grounds for rejection; sufficient grounds for sanction; sufficient grounds for retraction.
The journal will conduct a human and contextual analysis and will give authors the opportunity to respond before concluding that an infringement exists.
XV. On the Process Applicable to Statements Incompatible with the Effective Use of AI
When the journal identifies a possible omission, it may request clarification; correction of the statement; references; sources; previous versions; relevant prompts or outputs; original data; code; image files; and an explanation of the validation procedures. Allowed, ancillary, and correctable use that has been declared incompletely will not result in automatic rejection.
The editorial response will consider the nature of the use; extent; impact; reliability of the content; possibility of correction; existence of damage; intent; cooperation of the authors; and possible recidivism.
Depending on the severity, the journal may request supplementation of the statement; return the manuscript for revision; request a new verification of references; require correction of data, code, or images; temporarily suspend the review; restart the review stage; reject the manuscript; publish a correction; publish an editorial note; publish an expression of concern; retract the article; notify the appropriate institution or funder.
Fabrication, falsification, plagiarism, serious breach of confidentiality, misuse of personal data, creation of non‑existent references on a significant scale, or deliberate concealment may justify more severe measures.
Retraction will be reserved for situations in which the reliability, originality, or integrity of the article is materially compromised.
Before adopting a measure based on a possible violation, the authors will receive an objective description of the concern; an opportunity to provide explanations; a reasonable timeframe for response; the possibility to submit documents and records; and information about the decision made.
The authors may appeal when they identify a factual error; a procedural error; an inadequate interpretation of this Policy; or the existence of relevant evidence not considered.
The appeal will be reviewed, whenever possible, by a person who did not issue the original decision.
Any administrative sanctions by CNPq or other institutions will be decided by the respective bodies, not by the journal.
References
COMMITTEE ON PUBLICATION ETHICS (COPE). Artificial intelligence guidance. [n.p.]: COPE, [n.d.]. Available at: https://publicationethics.org/guidance. Accessed: Sep 5, 2026.https://publicationethics.org/guidance
NATIONAL COUNCIL FOR SCIENTIFIC AND TECHNOLOGICAL DEVELOPMENT (CNPq). Decree No. 2,664, of March 6, 2026. Establishes the Integrity Policy in CNPq’s Scientific Activity. Official Gazette of the Union: section 1, Brasília, DF, March 11, 2026, p. 4. Available at: CNPq Decree No. 2,664/2026. Accessed on: September 5, 2026.Portaria CNPq nº 2.664/2026
ELSEVIER. The use of generative AI and AI‑assisted technologies in writing for Elsevier. Amsterdam: Elsevier, [n.d.]. Available at: Elsevier — Generative AI in writing. Accessed on: Sep 5, 2026.Elsevier — Generative AI in writing
ELSEVIER. The use of generative AI and AI‑assisted technologies in the review process. Amsterdam: Elsevier, [n.d.]. Available at: Elsevier — Generative AI in the review process. Accessed on: 5 Sep. 2026.Elsevier — Generative AI in the review process
FRONTIERS. BE WISE: the six principles of human oversight in AI‑assisted research. Lausanne: Frontiers, 2026. Available at: Frontiers — BE WISE Framework. Accessed on: 5 Sep 2026.Frontiers — BE WISE Framework
MDPI. MDPI’s updated guidelines on artificial intelligence and authorship. Basel: MDPI, 20 Apr. 2023. Available at: MDPI — Updated Guidelines on Artificial Intelligence and Authorship. Accessed on: 5 Sep. 2026.MDPI — Updated Guidelines on Artificial Intelligence and Authorship
SPRINGER NATURE. AI guidance for researchers and communities. London: Springer Nature, 2026. Available at: Springer Nature — AI guidance for researchers and communities. Accessed on: Sep 5, 2026.Springer Nature — AI guidance for researchers and communities
TAYLOR & FRANCIS. AI policy. Abingdon: Taylor & Francis, [n.d.]. Available at: Taylor & Francis — AI Policy. Accessed on: Sep 5, 2026.Taylor & Francis — AI Policy
WILEY. Using AI tools in your research: a guide for research authors, editors, and reviewers. Hoboken: Wiley, 2025. Available at: Wiley — AI guidelines for researchers. Accessed on: Sep 5 2026.Wiley — AI guidelines for researchers
Artificial Intelligence Use Statement
In drafting this Policy, the generative artificial intelligence tool ChatGPT 5.6 Sol from OpenAI was used as an auxiliary resource for research and identification of normative sources and editorial guidelines, comparative review of journal and scientific publisher policies, systematization of the information found, and support for the organization and structuring of the text. The references, standards, and guidelines considered were subject to human consultation, selection, and evaluation, including, among others, CNPq Ordinance No. 2,664/2026 and recommendations and policies of national and international scientific organizations and publishers. The artificial intelligence was used as a support tool, not as an autonomous source of normative authority nor as responsible for editorial decisions. The definition of the adopted principles, the assessment of source relevance, the choice of incorporated rules, their suitability to the journal’s reality, and the final text review were carried out under human supervision and responsibility. Consequently, the content of this Policy was critically reviewed before publication, with responsibility for its drafting, interpretation, updating, and application remaining exclusively with the journal’s responsible team.