Policy on generative AI and AI-assisted tools

01

Purpose and scope

The policy applies to authors, reviewers and editors and covers AI-assisted writing, translation, analysis, coding, figures, literature support and peer review.

JGTEL recognises that large language models, chatbots, image generators and other AI-assisted tools may support parts of research and manuscript preparation.

Their use must not undermine originality, accuracy, confidentiality, transparency or trust in the scholarly record.

02

Human authorship and oversight

AI tools cannot be named as authors because they cannot take responsibility for a submission, approve the final version or answer questions about integrity.

• Human authors remain responsible for factual, legal and regulatory accuracy.
• Authors remain responsible for the originality and quality of analysis and argument.
• Every citation, quotation, dataset and authority must be checked against a reliable source.
• Confidential, personal or protected material must not be disclosed to an AI service without a lawful and authorised basis.
• AI output must be reviewed, verified and rewritten where necessary; it is a starting point, not an authority.

03

What must be disclosed

Material use of generative AI or an AI-assisted tool must be described specifically enough for editors and readers to understand its contribution.

1. Identify the tool and version where available.
2. State the purpose, such as language refinement, restructuring, translation, early drafting, code generation, data visualisation or source summarisation.
3. Describe the parts affected, the extent of the contribution and the human checks applied.

04

Potentially acceptable uses

Subject to disclosure and rigorous human verification, AI may assist with the following activities:

• Language refinement and translation that preserve the author's intended meaning
• Brainstorming, outlining or restructuring an early draft
• Code development, debugging or data visualisation where methods and outputs are independently checked
• Summarising literature to support discovery, provided authors read and cite the original sources
• Accessibility support and conversion between formats, with final human quality control

05

Prohibited and high-risk uses

• Fabricating, falsifying or selectively distorting evidence, results, quotations or legal authorities
• Submitting invented, unverifiable or hallucinated references
• Using AI output to conceal plagiarism or duplicate publication
• Replacing the author's core intellectual contribution or scholarly judgement
• Uploading confidential manuscripts, review reports, personal data or protected research material to an unauthorised service
• Generating deceptive images, data or analyses without transparent methods and validation

06

Reviewers and editors

Reviewers and editors must not upload a manuscript or any part of it to a public or third-party AI system unless the journal has expressly authorised a secure tool and the necessary confidentiality protections.

AI must not make the substantive recommendation or publication decision. Human reviewers and editors remain responsible for reading the work, evaluating evidence and reasoning, and writing their own reports and decisions.

Limited assistance with language or organisation may be used only where confidentiality is protected and the human participant checks and owns the final text.

07

Questions, investigation and record correction

Where AI use is unclear or potentially inappropriate, the journal may request an explanation, supporting material or revision. Serious or undisclosed misuse may result in rejection, correction, an expression of concern, retraction or referral to an institution, depending on the evidence and publication stage.

The policy will be reviewed periodically as technologies, research practices and professional standards develop.