AI Use Policy

Download

  1. Objective

Define clear strategies for the ethical, transparent and accountable use of Artificial Intelligence (AI) tools in the elaboration, submission, evaluation and publication of manuscripts on Cancer Control. The goal of the present policy is to ensure scientific integrity, data safety and compliance with national and international rules of publication.

  1. General Principles

The Revista Brasileira de Cancerologia:

  • Reaffirms its commitment with scientific and ethical integrity in oncologic research.
  • Recognizes that technological tools can support the scientific process.
  • Does not recognize AI systems as authors of manuscripts.
  • Requires transparency in the use of AI and supporting software (reference managers).
  • Complies with the recommendations of international bodies as:
    • Committee on Publication Ethics (Cope).
    • International Committee of Medical Journal Editors (ICMJE).
  1. Use of AI by authors

3.1 What is allowed

  • Grammar review and style refinement.
  • Translation.
  • Support in the elaboration of explanatory graphs, tables or figures[1] with declaration of use.
  • As tools to adjust references to the Vancouver style.

3.2 What is not allowed

  • Creation of text.
  • Structural organization of the text.
  • Production or editing of clinical images or of human beings.
  • Analysis, interpretation of data and results or conclusion about any aspect of the research.
  • Creation of references.
  • As bibliographic source/reference.

The declaration of use of AI must be included in:

  • Methods (when applicable).
  • In the specific topic “DECLARATION OF AI USE”.

The declaration must include the following language (mandatory):

  • For what was used, “Text revision”, “Translation”, “Norms”, ...
  • Version of AI used: “ChatGPT-4”, “ChatGPT-4.5”.
  • Prompt, instruction, question or command used in PDF and snapshot (static copy) of your message (link) and date.

3.3 Responsibility

  • The authors are fully responsible for the content.
  • AI must not be cited as author.
  • AI is not authorized to fabricate, manipulate of omit data.
  • Generation of inexistent references is deemed as a grave error which may lead to the potential rejection of the manuscript regardless of the editorial phase (submission, analysis, approval and editing).
  1. Use of AI in peer-review
  • Manuscripts submitted for evaluation must not be inserted into AI public tools that store or utilize data for training.
  • Reviewers must preserve anonymity and confidentiality.
  • The editor must be informed whether AI was utilized as supporting tool (organization of comments).

Editorial decisions must not be made exclusively by automated systems.

  1. Editorial Use of AI
  • Initial triage of manuscripts.
  • Verification of textual similarity.
  • Support to detect formal inconsistencies

Final decision is human-based and by a collegiate body.

  1. Use of Reference Managers

RBC allows the use of the following Reference Managers reported in the submission form and declaration of copyright:

  • EndNote.
  • Zotero.
  • Mendeley.
  • The authors must review all the references manually prior to the submission.
  • The authors must remove the hyperlinks from the citations (progressive numeration, tables and/or charts) created by reference managers).
  • Mandatory verification of:
  • Authors.
  • Title.
  • Journal.
  • Publication data (volume, number and page/e-location, publisher, year, place of publication etc.).
  • Date of access (when applicable).
  • Link/DOI.

After the approval of the manuscript, references will NOT be replaced, unless a formal justification is submitted for editor’s review.

6.1 Responsibility

The use of managers does not exempt authors from their responsibility for:

  • Incorrect/inexistent citations.
  • Incorrect/inexistent references.
  • Inclusion of inexistent articles.
  • Formatting errors in disagreement with RBC norms.
  1. Policy update

This policy will be updated periodically to incorporate:

  • New advances of AI applied to health.
  • Updates of guidelines and international/national regulatory rules.

[1]Figures, according to the definition of the “Instituto Brasileiro de Geografia e Estatística (IBGE)”, comprises photographs, however, RBC will not accept AI-produced photographs and clinical images or of human beings. Only visual representations (diagrams, infographics, dashboards) that do not compromise the distinction between AI-assisted or human-produced will be accepted.