Artificial intelligence (AI) is transforming some processes in research and publishing, and its use continues to expand. However, it is widely accepted that it can lead, consciously or unconsciously, to poor practices within the author-reviewer-editor triad. Furthermore, AI can potentially modify or replace the critical thinking skills of authors and the expert judgment of reviewers, compromising the quality of publications.

At Sociedad y Ambiente, we recognize the importance of reporting the use of AI, whether it was secondary or intensive, in submissions and publications. This is a preventative strategy that is the responsibility of the editorial team and helps minimize the spread of risks and reduce bias, in accordance with the Heredia Declaration.

For the reasons stated above, we have prepared a series of policies for authors and reviewers, and we have also disclosed the processing of manuscripts using AI-assisted programs, with the aim of making their use in research, manuscript preparation, peer review, and editorial work more transparent. Our goal is to provide greater reliability throughout the knowledge production chain and in our publications. These policies are described below:

Authors who use AI are responsible for verifying the accuracy of the information and ensuring that the intellectual property of the consulted sources is not infringed.

AI should not be used for text generation, theoretical reflection, discussion, or conclusions in manuscripts, which are strictly personal or collaborative exercises.

Manuscripts that exceed one-third of their content in AI use may be withdrawn during the technical review stage at the editorial team's discretion.

All use of AI must be reported in the comments to the editor, as indicated in the checklist, and in the manuscript methodology, according to the instructions for its preparation.

Reviewers should not use AI in the manuscript review process, as the purpose of the invitation is to obtain a specialized opinion on the content of the proposal, based on their knowledge and experience.

The editorial team may use AI-assisted programs to detect the percentage of similarity with other bibliographic sources and the amount of text potentially generated using this type of models. This ensures the originality of the submissions and the accuracy of the declaration regarding the use of AI, as requested in the checklist.

If AI is implemented in other ways during the editorial process, this will be detailed in these policies and communicated to authors during the submission stage to obtain their consent via the checklist.