Artificial intelligence policy

Investigación en Discapacidad is committed to upholding the highest standards of scientific integrity, transparency, and inclusivity, particularly in research involving disability. As generative AI and AI-assisted tools become ubiquitous in research and publishing, this policy defines acceptable and prohibited uses for authors, peer reviewers, and editors when handling manuscripts, text, images, data, or any other file type submitted to the journal. The policy follows the Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly Work in Medical Journals from the International Committee of Medical Journal Editors (ICMJE) and evolving best practices for responsible AI use in scholarly publishing.

1. General Principles for All Parties

  • Transparency: Any use of AI tools must be disclosed explicitly.

  • Accountability: Human authors, reviewers, and editors remain fully responsible for all content, decisions, and outputs.

  • Confidentiality: AI tools must not be fed with unpublished manuscripts, review data, or editorial correspondence.

  • Inclusivity: AI use must not introduce or amplify bias, especially against persons with disabilities or other vulnerable groups.

2. Guidelines for Authors

Authors may use AI tools to improve language, readability, or data analysis, subject to strict rules.

2.1 Permitted Uses (with disclosure)

  • Language polishing, grammar correction, or paraphrasing.

  • Statistical analysis or code generation for reproducible research.

  • Image enhancement (e.g., brightness, contrast) – but no content manipulation.

  • Formatting references or generating bibliographies.

2.2 Prohibited Uses

  • No AI-generated content as core scientific contribution (hypotheses, conclusions, original findings).

  • No undisclosed AI use – any assistance must be declared.

  • No AI as author – AI tools cannot be listed as co-authors nor be credited for authorship because they cannot take responsibility for the work.

  • No image creation or manipulation that alters scientific meaning (e.g., generating fake Western blots, adding/removing lesions in clinical images, fabricating radiological findings).

  • No AI-based plagiarism or paraphrasing to circumvent similarity checks.

2.3 Disclosure Statement (Mandatory)

Authors must include a clear statement in the Methods or Acknowledgments section (or in a separate section at the end of the manuscript) describing:

  • Name of the AI tool (including version).

  • Purpose of use (e.g., language editing, data visualization, code help).

  • Date(s) of use.

  • Confirmation of human review and responsibility.

Example:
“During the preparation of this work, the authors used [Tool name, version] for [specific task, e.g., grammar correction of the abstract]. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the final publication.”

If no AI was used, a simple statement: “No AI tools were used in the creation of this manuscript.”

2.4 Special Requirements for Images and Figures

  • For any AI-assisted image processing (e.g., deblurring, segmentation), authors must submit the original unprocessed image and a log of all steps.

  • AI-generated synthetic images (e.g., diagrams, illustrations) are allowed only if clearly labeled as “AI-generated” and their purpose is educational (not as primary data).

  • Deepfakes or AI-generated patient images are strictly prohibited.

3. Guidelines for Reviewers

Reviewers’ confidentiality obligations supersede any convenience of AI tools.

3.1 Prohibited Uses for Reviewers

  • Do not use AI to draft the review report, as it may generate incorrect, biased, or superficial critiques and violates confidentiality.

  • Do not rely on AI to verify scientific claims or references; you must do so manually.

3.2 Permitted Uses (with caution)

  • Using publicly available AI search engines to check published literature, provided no manuscript details are entered.

  • Grammar/spell checkers that do not store or send data to third-party servers (local software).

3.3 Disclosure

If a reviewer inadvertently used a non-internet connected AI tool (e.g., offline grammar checker), no formal disclosure is needed. However, any use of generative AI to assist with writing the review must be disclosed to the editor and may lead to disqualification.

4. Guidelines for Editors

Editors must maintain human decision‑making at all stages.

4.1 Permitted Uses

  • Using AI-based plagiarism detection or image duplication software

  • Using AI to screen for potential reviewers (via public databases) without sharing manuscript details.

  • Using grammar/style checkers for editorial correspondence.

4.2 Prohibited Uses

  • No AI to make final acceptance or rejection decisions.

  • No AI to generate editorial decisions or letters without substantive human review.

  • No AI to override peer review evaluations.

  • No AI to classify manuscripts into ethical or scientific quality categories automatically.

4.3 Disclosure and Transparency

Editors must declare any use of AI in the editorial workflow in the journal’s annual report. For specific cases (e.g., image fraud detection), authors will be informed.

4.4 Responsibility

The handling editor is always fully responsible for the final decision and for ensuring that no AI tool introduced bias (e.g., against authors from low-resource settings, non-native English speakers, or disability topics).

5. Special Considerations for All File Types

File type AI use permitted (with disclosure) AI use prohibited
Text (manuscript, reviews, letters) Language polishing, summarization of own draft Generating original scientific claims, undisclosed writing, full review generation
Images (clinical, radiological, histology, blots) Noise reduction, background correction (log required) Creating, adding, removing, or duplicating features; generating synthetic patient images
Data files (Excel, CSV, SPSS) Code generation for analysis (Python, R) Fabricating data, auto‑completing missing values without justification
Video/audio (rehabilitation procedures, interviews) Speech-to-text for transcripts (if consented) Generating fake movements, altering patient identity beyond anonymization
Supplementary files Language translation of captions Using AI to create fake raw data

6. Bias and Accessibility Statement

Because Investigación en Discapacidad serves the disability community, all AI use must:

  • Not perpetuate ableist biases (e.g., AI image generators that default to non-disabled bodies).

  • Not discriminate against authors using AI to improve accessibility (e.g., screen-reader friendly formatting).

  • Be reported if an AI tool misrepresents disability-related concepts (editors will issue a correction).

7. Monitoring and Enforcement

  • Plagiarism check: All submissions run through similarity checks (includes AI-generated text detection where possible).

  • Image screening: Suspicious images will be examined using AI detection tools.

  • Reviewer audits: Editors may randomly ask reviewers to confirm no AI was used on confidential material.

8. Changes to This Policy

This policy will be reviewed annually. The journal will notify all registered users of major changes via email and a notice on the journal website (https://dsm.inr.gob.mx/indiscap/index.php/INDISCAP/index).