Generative AI for Reviewer

This page summarises the obligations of peer reviewers of Advance Sustainable Science, Engineering and Technology (ASSET) regarding generative artificial intelligence (GenAI) and AI-assisted technologies during the peer-review process. It is derived from, and must be read together with, the full ASSET Generative AI Policy, which prevails in case of any difference.

1. Confidentiality

Peer review is a confidential scholarly activity. Reviewers must not upload, paste, submit, or otherwise transfer any of the following to an external GenAI system:

  • the manuscript or substantial portions of it;
  • unpublished data, figures, tables, or supplementary materials;
  • reviewer or editorial correspondence;
  • any other confidential information obtained through peer review.

This applies even if an AI service is described as "free", "private", or "secure". Reviewers should use only AI tools explicitly authorised by ASSET for the relevant purpose.

2. Substantive Evaluation

Reviewers must personally evaluate the manuscript and must not delegate substantive assessment to GenAI. AI must not determine:

  • whether the manuscript should be accepted or rejected;
  • its novelty, methodological appropriateness, or validity;
  • its overall scholarly quality;
  • the substantive recommendations made to the editor.

3. Limited Use for Review Reports

Reviewers may use AI-assisted tools for limited language or structural improvement of their own review reports, provided that:

  • no confidential manuscript material is entered into the AI system;
  • the reviewer's independent scholarly judgment remains the basis of the review;
  • the reviewer verifies and edits the output;
  • the use is declared where required by the policy.

4. Reviewer Responsibility

Reviewers remain fully responsible for the content and judgment expressed in their review reports, including any AI-assisted wording.

5. Raising Concerns

If a reviewer suspects undisclosed or inappropriate AI use in a manuscript (e.g., fabricated references or data, misleading images, or a missing or inaccurate GenAI declaration), the reviewer should inform the handling editor in the confidential comments, explaining the basis for the concern.

See also: Generative AI Policy | Generative AI for Author | Generative AI for Editor