AI Use Policy

EDITORIAL POLICY ON THE USE OF ARTIFICIAL INTELLIGENCE
Aligned with CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico – National Council for Scientific and Technological Development, Brazil) Ordinance No. 2,664/2026 and the practices of leading international scientific publishers

1. Introduction and Purpose

Generative artificial intelligence (AI) has significantly transformed the workflows of scientific research. Tools such as large language models (LLMs) are widely used by authors and reviewers to assist in writing, synthesis, grammatical correction, and data analysis. This scenario requires the Revista Brasileira de Geomorfologia (RBGeomorfologia) to establish clear guidelines to ensure integrity, transparency, and responsibility in the production and evaluation of published scientific knowledge.

This policy is based on two main pillars:

  • CNPq Ordinance No. 2,664, of March 6, 2026, which establishes the Policy on Integrity in Scientific Activity of CNPq and sets specific guidelines on the use of Generative Artificial Intelligence in research and scientific publishing (Art. 9, item I, subitems c, d, e, f);
  • The editorial practices of leading international scientific publishers and the recommendations of the Committee on Publication Ethics (COPE).

The fundamental principle guiding this policy, in full compliance with CNPq Ordinance No. 2,664/2026, is non-negotiable human responsibility: AI is a support tool, and all ethical, scientific, and legal responsibility for the published content, including any plagiarism or inaccuracies generated by generative AI, rests exclusively with the human authors.

Mandatory legal basis - CNPq Ordinance No. 2,664/2026

Art. 9, I, c) declare the use of Generative Artificial Intelligence (GAI) tools, of any type and at any stage of research development...

Art. 9, I, d) submission of content generated by GAI as if it were of human authorship is prohibited...

Art. 9, I, e) insertion of third-party research projects into GAI tools for the preparation of scientific reviews is prohibited...

Art. 9, I, f) assume full responsibility for the final content of the research...

 

2. Authorship and AI Tools

2.1 Prohibition of AI Authorship

Generative AI tools, including large-scale language models (LLMs) such as ChatGPT, Gemini, Claude, Copilot, DeepSeek, and similar, CANNOT be listed as authors or co-authors of any manuscript submitted to RBGeomorfologia. Scientific authorship requires the ability to assume ethical and legal responsibility for the published content.

This prohibition is in accordance with: (a) CNPq Ordinance No. 2,664/2026, Art. 9, I, d, which prohibits the submission of content generated by generative AI as if it were of human authorship; and (b) the unanimous position of the Committee on Publication Ethics (COPE) and leading international scientific publishers.

2.2 Full Responsibility of the Authors

In compliance with Art. 9, I, subitems d and f of CNPq Ordinance No. 2,664/2026, authors are fully responsible for the final content of the manuscript, including any plagiarism, inaccuracies, hallucinations, or distortions generated by generative AI tools used at any stage of the research. This includes:

  • Individual verification of each bibliographic reference generated or suggested by AI against primary sources;
  • Validation of all data, analyses, and results processed with AI support;
  • Critical review of all text, translations, or syntheses produced with generative AI support;
  • Ensuring that no AI-generated, fabricated, or inaccurate information is included in the published manuscript.

2.3 Obligation to Declare at Any Stage of the Research

In compliance with Art. 9, I, c, of CNPq Ordinance No. 2,664/2026, the use of generative AI tools must be declared at any stage of the research development, including:

  • Conception phase: use of generative AI to identify gaps in the literature, formulate hypotheses, or define the scope of the research;
  • Writing phase: use of generative AI to draft, review, translate, or improve the manuscript text;
  • Data analysis phase: use of generative AI to process, analyze, or visualize data;
  • Submission phase: use of generative AI in preparing the cover letter or responses to reviewers.

Basic spell-checking tools integrated into word processors (such as the Microsoft Word spell checker) are exempt from this requirement.

The declaration must be included in a specific section of the manuscript, immediately before the References, under the title "Declaration of Artificial Intelligence Use", specifying: (a) the name of the tool used; (b) the stage and purpose of use; and (c) the extent of human supervision and verification performed.

2.4 Uses of AI by Authors

RBGeomorfologia recognizes that generative AI tools may contribute to the scientific production process when used transparently and under supervision. The table below summarizes permitted and non-permitted uses:

PERMITTED uses (with declaration) NON-PERMITTED uses
Grammatical, spelling, and stylistic revision of the text Listing AI tools as authors or co-authors
Improvement of fluency and readability Submitting AI-generated content as if human, without declaration
Translation of manuscripts (with human review) Generating results, conclusions, or discussions without human validation
Support for literature research (with source verification) Modifying results, images, or graphs to favor hypotheses
Support for statistical analysis and data processing Fabricating or altering research data
Visualization of real data (with declaration) Uploading unpublished manuscripts to AI without risk assessment
Generation of illustrative figures (not based on primary data) Generating bibliographic references without verification in primary sources

3. Use of AI by Reviewers

3.1 Prohibition of Sharing Manuscripts with Generative AI

In direct compliance with Art. 9, I, e of CNPq Ordinance No. 2,664/2026 — which prohibits "the insertion of third-party research projects into generative AI tools for the preparation of scientific reviews" — it is PROHIBITED to upload, insert, copy, or share manuscripts under review, or any part thereof (texts, tables, figures, data), into generative AI tools of any kind.

This prohibition aims to protect:

  • the confidentiality of the peer review process;
  • the authors' rights over unpublished data and results;
  • the integrity of the evaluation process.

3.2 Permitted Uses of AI by Reviewers

The following auxiliary uses are permitted, provided that the manuscript content is not shared with external tools:

  • Use of generative AI to improve the grammar and clarity of the review report itself (after being fully written by the reviewer);
  • Consultation of scientific databases and search tools with AI support to verify factual information in the public domain.

3.3 Responsibility and Declaration for Reviewers

Technical judgment, editorial recommendations, and the evaluation of the scientific merit of the work must be exclusively of human authorship. If the reviewer uses generative AI in preparing their review report (for linguistic revision purposes only), it is not necessary to inform the handling editor at the time of submission.

RBGeomorfologia reserves the right not to consider reviews that are evidently generated automatically by generative AI, without critical and specialized analysis of the scientific content.

4. Images, Figures, and Data

The visual integrity of data is fundamental to geomorphological science. The following rules apply specifically to the use of AI in visual representations:

  • It is PROHIBITED to use AI in scientific images to alter the informational content of visual representations in a way that distorts, conceals, or favors research results, which constitutes falsification under Art. 5, XII and Art. 33, I of CNPq Ordinance No. 2,664/2026;
  • Basic brightness and contrast adjustments are permitted provided they are applied uniformly and do not alter data interpretation;
  • The use of AI for visualization and graphical representation of real data (such as plots generated by analysis scripts) is permitted;
  • Schematic diagrams and illustrative figures (not representing primary data) may be produced with generative AI support, provided they are identified as such in the captions.

5. Declaration Templates

5.1 Template A (Use of AI in writing or text revision)

Declaration of Artificial Intelligence Use

During the preparation of this manuscript, the authors used the tool [TOOL NAME, e.g., ChatGPT (OpenAI, GPT-4o version)] for the purpose of [PURPOSE, e.g., improving grammar and text fluency / assisting in translation into English]. After using the tool, the authors critically reviewed and edited all content, assuming full and complete responsibility for the published manuscript, including any inaccuracies or errors.

5.2 Template B — (Use of AI in data analysis or methodology)

Declaration of Artificial Intelligence Use

In the [STAGE, e.g., statistical analysis / image processing / literature review] phase of this study, the authors used the tool [NAME, e.g., GitHub Copilot / ChatGPT] for the purpose of [DETAILED PURPOSE]. All results were verified and validated by the authors. The original data and scripts are available at [REPOSITORY/LINK], ensuring reproducibility. The authors assume full responsibility for the content.

5.3 Template C (Use of AI in research conception)

Declaration of Artificial Intelligence Use

In the conception phase of this research, the authors used the tool [NAME] to [PURPOSE, e.g., identify gaps in the literature / assist in defining the research problem]. The hypotheses, objectives, and final scope of the research were defined exclusively by the authors, who assume full responsibility for the scientific content.

5.4 Template D (No AI Use)

Declaration of Artificial Intelligence Use

The authors declare that no generative artificial intelligence tools were used at any stage of the development of this manuscript (conception, writing, data analysis, or submission).

6. Compliance, Violations, and Consequences

Non-compliance with this policy constitutes a potential violation of scientific integrity under the terms of CNPq Ordinance No. 2,664/2026. Depending on the severity, the infraction may be classified as serious or very serious (Art. 32 and 33 of the Ordinance), subjecting the researcher to the sanctions provided in Art. 34, including suspension of CNPq fellowships and grants, prohibition from participating in funding actions, and reimbursement of funds.

Within the editorial scope of RBGeomorfologia, consequences may include:

  • Return of the manuscript for correction before review;
  • Rejection of the manuscript during the review process;
  • Retraction of the article after publication, in cases of serious violation;
  • Notification to CNPq and the authors' home institution(s), when applicable.

RBGeomorfologia may use technological tools to assist in detecting undeclared use of generative AI and inconsistencies in scientific images, always under the supervision of the editorial team.

7. Legal Basis and References

7.1 National Regulation

  • CNPq Ordinance No. 2,664, of March 6, 2026 - Policy on Integrity in Scientific Activity of CNPq (Art. 9, I, c, d, e, f). Published in the Official Gazette (DOU) of March 11, 2026, Section 1, p. 4.

7.2 International Guidelines