AI Ethics and Procedure
AI and AI-assisted technologies may support scholarly work, but they must not replace the intellectual responsibility, critical judgment, scientific reasoning, or accountability of human researchers.
Authors remain fully responsible for the accuracy, originality, integrity, validity, and reproducibility of all content submitted to SEMPOA, regardless of whether AI-assisted technologies were used.
The journal evaluates manuscripts on the basis of scholarly quality, originality, methodological or mathematical rigor, transparency, ethical compliance, and relevance to the journal's scope rather than solely on whether AI tools were used.
Authors may use AI or AI-assisted technologies where such use is appropriate, lawful, scientifically justified, and subject to human supervision.
Examples may include:
- language improvement and grammatical correction;
- formatting and readability assistance;
- computer programming or code assistance;
- mathematical or symbolic computation;
- data processing and statistical analysis;
- algorithm development or implementation;
- visualization assistance;
- AI-based educational analysis;
- machine-learning or artificial-intelligence methods used as part of the research methodology.
When AI itself forms part of the research method, its use should be described with sufficient methodological detail to permit scientific evaluation and, where feasible, reproducibility.
Authors must disclose the use of generative AI or other AI-assisted technologies when such technologies have made a material contribution to manuscript preparation or the research process.
Routine use limited to spelling correction, grammar checking, basic formatting, or similar conventional language-editing functions does not normally require a detailed declaration.
Where disclosure is required, authors should identify, where applicable:
- the name of the AI system or software;
- the provider or developer;
- the version or model, where known;
- the purpose for which the tool was used;
- the part of the research or manuscript affected;
- how the AI-generated output was verified by the authors.
"During the preparation of this manuscript, the authors used [tool/model name] for [purpose]. All outputs were reviewed, verified, and revised by the authors, who take full responsibility for the final content of the manuscript."
AI systems, large language models, chatbots, software applications, algorithms, and other automated technologies cannot be listed as authors or co-authors. Authorship requires accountability and responsibilities that can only be assumed by human contributors.
Generative AI systems must not be treated as authoritative scholarly sources and should not be cited as substitutes for original scientific publications.
Authors must independently verify all factual statements, mathematical expressions, references, quotations, numerical results, bibliographic information, and other outputs obtained through AI-assisted systems.
Authors must consult and cite the original scholarly source rather than citing an AI-generated summary of that source.
Authors are responsible for verifying every reference in the manuscript against the original publication.
AI-generated or AI-suggested references must not be included without verification.
Fabricated references, nonexistent DOIs, incorrect author names, invented publication details, or citations to sources that the authors have not verified may constitute a research-integrity concern.
Where AI tools are used to assist with mathematical derivations, proofs, symbolic computation, theorem exploration, algorithm development, numerical experimentation, or optimization, authors remain responsible for independently establishing the validity of the resulting mathematical claims.
AI-generated proofs or derivations must not be accepted by the authors without mathematical verification.
When AI-assisted computation materially contributes to the findings, authors should provide sufficient details concerning the computational environment, algorithms, parameter settings, software, and validation procedures where appropriate.
Research involving AI-supported mathematics teaching, intelligent tutoring systems, learning analytics, automated assessment, educational chatbots, or other AI-based educational technologies must explain the role of AI within the study.
Researchers must carefully consider potential bias, fairness, accessibility, privacy, and the possible effects of automated decision-making on learners and educators.
Personally identifiable or confidential student, teacher, or participant information must not be uploaded to public or third-party AI systems unless such processing is ethically approved, legally permitted, appropriately protected, and consistent with informed consent and applicable data-protection requirements.
The use of AI-generated, simulated, or synthetic data must be clearly identified and scientifically justified.
Synthetic or AI-generated observations must not be presented as empirical observations collected from real participants, institutions, experiments, or naturally occurring datasets.
Authors should explain how synthetic data were generated and validated and should clearly distinguish them from real-world data throughout the manuscript.
Authors must disclose material use of generative AI in producing or substantially modifying figures, illustrations, diagrams, or other visual research content.
AI-generated visual material must not falsify, fabricate, obscure, or misleadingly alter research evidence. Authors are responsible for ensuring that all published visual content accurately represents the underlying research.
The following practices are not permitted:
- listing an AI system as an author;
- fabricating or manipulating research data using AI;
- inventing participants, experiments, survey responses, or observations;
- creating false or nonexistent references;
- presenting AI-generated mathematical results as verified results without human validation;
- using AI to conceal plagiarism or inappropriate text reuse;
- using AI to manipulate figures or research evidence deceptively;
- submitting substantially AI-generated scholarly content without appropriate human intellectual contribution and verification;
- uploading confidential manuscript or participant information into AI systems without appropriate authorization.
SEMPOA may use automated tools, including AI-writing indicators or research-integrity screening systems, as part of the editorial assessment process.
Results produced by automated detection systems are treated as screening indicators rather than conclusive evidence of misconduct.
The journal does not impose an automatic acceptance or rejection decision solely on the basis of an AI-detection percentage.
Any concern identified by an automated system will be subject to appropriate human editorial assessment before an editorial or research-integrity decision is made.
Manuscripts submitted to SEMPOA are confidential documents.
Reviewers must not upload manuscripts, supplementary files, unpublished data, or confidential editorial information to public generative AI systems or other services that may retain, process, reproduce, or use the submitted material.
Peer reviewers must not use generative AI to create substantive peer-review assessments or recommendations.
Review reports must reflect the reviewer's own scholarly evaluation, expertise, reasoning, and judgment.
Editors must protect the confidentiality of submitted manuscripts and associated editorial information.
Editors must not upload confidential manuscript content into public generative AI systems or use generative AI to independently generate editorial assessments or publication decisions.
Final editorial decisions must always be made by qualified human editors based on the manuscript, peer-review reports, journal policies, and scholarly judgment.
The journal may use automated systems to assist with tasks such as similarity screening, reference verification, metadata validation, integrity checks, or detection of potentially problematic manuscript content.
Such tools support but do not replace editorial judgment. Significant findings produced by automated systems must be evaluated by an editor or appropriately qualified member of the journal's editorial team.
When the editorial team identifies reasonable concerns regarding undisclosed, inappropriate, or potentially unethical AI use, the following procedure may be applied:
Step 1 — Preliminary Assessment
The editor reviews the manuscript and available evidence.
Step 2 — Author Clarification
The corresponding author may be asked to explain the use of AI tools and provide supporting information.
Step 3 — Verification
Editors evaluate the author's explanation together with the manuscript, supporting materials, and relevant screening results.
Step 4 — Corrective Action
Authors may be required to revise the manuscript, add an AI-use declaration, correct unsupported information, provide underlying data or calculations, or otherwise address the identified concern.
Step 5 — Editorial Decision
Serious or unresolved breaches may result in rejection, withdrawal of consideration, or post-publication action where appropriate.
SEMPOA does not use a fixed percentage of AI-detected text as an automatic criterion for determining research misconduct, rejection, or acceptance.
Automated AI-detection results may assist editorial screening, but decisions are based on human evaluation of transparency, authorship responsibility, scholarly contribution, research validity, originality, and compliance with journal policies.
Regardless of the technology used in conducting research or preparing a manuscript, human authors retain full responsibility for:
- the originality of the submitted work;
- the accuracy of data and mathematical results;
- the validity of references and citations;
- the interpretation of findings;
- ethical compliance;
- the integrity of figures and tables;
- the disclosure of AI-assisted activities where required;
- and the complete content of the final manuscript.
