Authors are responsible for all material submitted to GGSR, including text, references, images, data processing, and code produced with automated assistance. AI systems cannot be authors because they cannot approve publication, make accountable declarations, or accept responsibility for the research.

Disclosure of substantive use

Disclose generative-AI use that substantively contributes to drafting, interpretation, analysis, code, images, or other research outputs. Identify the tool and version where available, the purpose and extent of use, and the checks performed by the authors. Place research-method uses in the methods section and other substantive uses in an appropriate declaration. Routine spelling or grammar correction that does not generate substantive content ordinarily does not require a separate statement.

An informative declaration might explain that a named tool assisted with a specified coding task and that the authors inspected the code and tested its outputs. A vague statement that “AI was used” does not explain the tool’s role or how reliability was assessed.

AI as part of a research design

When an AI model is used to classify documents, generate measures, analyze discourse, or otherwise produce research evidence, describe the input material, model or service, relevant dates and settings, prompting or training procedure, validation, and limitations. Explain the implications of model changes, nondeterministic outputs, language coverage, and potential bias for the reported findings. Document procedures sufficiently for meaningful assessment, subject to rights and confidentiality constraints.

Verification and unacceptable practices

Authors must verify factual statements, quotations, references, calculations, and other outputs against appropriate sources or tests. Fabricated citations, invented interviews, synthetic observations presented as collected data, and misleading alterations of evidence are unacceptable. AI-generated assistance does not excuse plagiarism, inaccurate attribution, or unreported manipulation.

Clearly identify synthetic or generated material when it is a legitimate object or tool of the research. Do not represent generated images, documents, or quotations as authentic source evidence.

Confidentiality and editorial use

Authors, editors, and reviewers must not transmit confidential manuscripts, personal information, unpublished evidence, or review correspondence to an AI service without appropriate authorization and safeguards. Public availability of a tool is not evidence that it protects submitted material.

Editors and reviewers remain responsible for their own assessment and must not delegate a review or editorial decision to an automated system. Any permitted assistance must preserve confidentiality and independent judgement. See Research Integrity for the journal’s expectations for evidence and reporting.