AI Policy

Introduction
Fracture and Structural Integrity recognizes the potential of Artificial Intelligence (AI) to enhance research. In alignment with the TITAN2025 Guidelines, this policy ensures transparency, human oversight, and technical robustess. The Editors will evaluate each case to prevent misuse and ensure that AI application does not compromise the scientific record.

General Principles
Transparency & traceability: In accordance with TITAN2025, every AI interaction that significantly impacts the research output must be documented and traceable.
Human-centric integrity: Accountability remains exclusively with human authors. AI cannot be held liable; therefore, it cannot be a legal entity in the publishing process.
Data sovereignty & privacy: AI tools must comply with high standards of data security, ensuring that "training data" does not infringe upon intellectual property or personal privacy.
Algorithmic bias mitigation: Users must actively monitor and correct any bias or "hallucinations" inherent in generative models.

For Authors
Permitted uses of AI: Authors may use generative AI for idea generation, language polishing, and coding. While basic grammar correction (e.g., spellcheck) does not require disclosure, any structural or substantive AI intervention must be declared.
Required disclosure (TITAN2025 Standard): Any use of generative AI in text, image, or data generation must be disclosed in a dedicated "AI Statement" (within Methods or Acknowledgments). This must include:
a) Model identification: name, version, and developer (e.g., GPT-4o, Claude 3.5).
b) Functional scope: how the tool was used (e.g., "summarization of raw sensor data").
c) Validation method: how the author verified the AI output for accuracy.
Accountability and prohibited uses
a) Authorship: AI tools cannot be listed as authors.
b) Data integrity: AI must not be used to fabricate or "smooth" original research data (e.g., stress-strain curves or micrograph features).
c) Image integrity: Enhancing, moving, or removing features in scientific images via AI is strictly prohibited unless the AI application itself is the subject of the study.

Authors must download, complete, sign, and upload the AI Declaration during submission.

Failure to declare the use of Artificial Intelligence (AI) in the preparation of the manuscript and failure to upload the mandatory AI Declaration form, will result in the automatic rejection of the paper. Please note that all submissions will be rigorously screened for undisclosed AI usage using iThenticate.

For Reviewers
Confidentiality & "No-Upload" Rule
Under TITAN2025, the confidentiality of the peer review is paramount.
a) Prohibition: Reviewers are strictly forbidden from uploading manuscripts (or parts of them) into generative AI tools.
c) Risk: Uploading content often feeds the AI’s training model, resulting in a breach of the author's intellectual property.
Original assessment
Peer review requires human critical thinking. AI may be used to refine the language of a review, but the critical evaluation and the final recommendation must be purely human-derived.

For Editors
Human oversight (the "Human-in-the-Loop" principle). Editors maintain final responsibility. AI should not be used to:
a) Generate final decision letters.
b) Make acceptance/rejection decisions based on automated summaries.
Permitted Uses
Editors may utilise AI-driven tools for reviewer discovery and plagiarism/image manipulation detection, provided these tools comply with TITAN2025 data protection standards.

Editorial Investigation
In cases of suspected undisclosed AI use, the journal will follow COPE (Committee on Publication Ethics) guidelines. Failure to disclose AI use in accordance with TITAN2025 standards may result in immediate rejection or retraction of the work.

Note on TITAN2025: This update specifically emphasises the validation process. Authors are now explicitly required to state how they checked the AI's work, rather than merely stating that they used it.