Policy on the Use of Artificial Intelligence

The Belize Journal of Medicine (BJM) acknowledges the increasing use of artificial intelligence (AI) tools, including generative language models, machine learning systems, and image-generation tools, in academic and scientific work. The journal adopts a permissive policy with restrictions, grounded in the principles of human responsibility, transparency, scientific integrity, and confidentiality.

1. General Principles

In accordance with the ICMJE Recommendations

  • AI tools may not be listed as authors or co-authors.
  • Full responsibility for the content rests solely with human authors.
  • Any use of AI must be transparently disclosed.

Authorship entails intellectual and ethical accountability, which cannot be assumed by automated systems.

2. Permitted Use of AI by Authors

AI tools may be used for language editing and translation, preliminary structuring or outlining of ideas, statistical programming support (when disclosed), exploratory literature searches, data modeling or content generation when AI is explicitly part of the study design.

All AI-assisted content must undergo critical human review.

3. Mandatory Disclosure of AI Use

If AI tools were used, authors must disclose:

  • Name of the tool.
  • Version (if applicable).
  • Date of access.
  • Specific purpose of use.

Disclosure must appear:

  • In the cover letter.
  • In the manuscript (n the Acknowledgments section, if used for writing assistance, in the Methods section (if used for data analysis or figure generation).

Authors must carefully verify all AI-generated content, as such tools may produce inaccurate or biased outputs.

4. Citation of AI Tools

If cited, AI tools must be referenced according to Vancouver style for electronic resources, following guidance https://www.icmje.org/recommendations/browse/artificial-intelligence/

AI systems must not be cited as authors.

5. Prohibited Uses

The following are not permitted:

  • Listing AI as an author.
  • Delegating scientific interpretation entirely to AI.
  • Generating references without human verification.
  • Fabricating or manipulating data.
  • Creating clinical images without disclosure.
  • Submitting AI-generated content without substantial human revision.

6. Use of AI by Reviewers

Reviewers must not:

  • Upload unpublished manuscripts to public AI systems.
  • Delegate peer review reports to AI tools.

They may use language-editing tools provided confidentiality is preserved.

7. Use of AI by Editors

Editors must not:

  • Replace editorial judgment with automated systems.
  • Upload manuscripts to public AI platforms.
  • Rely solely on AI-detection software for editorial decisions.

8. Detection and Editorial Assessment

BJM acknowledges that AI-detection systems are unreliable and may produce false positives

Editorial decisions will be based on comprehensive scientific evaluation rather than automated thresholds.

9. Non-Compliance

Failure to comply with this policy may result in: Requests for clarification; Additional review; Rejection; Retraction; Institutional notification in cases of misconduct.