Generative AI and AI-Assisted Technologies Policy

This policy governs the use of generative artificial intelligence and AI-assisted technologies in research, manuscript preparation, peer review and editorial work for the journal. It applies to tools that generate or materially transform text, images, audio, code, data, analyses or recommendations. The policy supports responsible assistance while preserving human authorship, accountability, confidentiality and the reliability of the scholarly record. English is the reference version; Uzbek and Russian versions express the same policy.

AI technologies change rapidly. The principles below apply regardless of product or vendor. The journal does not claim that an AI detector can establish authorship or misconduct, that a particular AI service is approved, or that confidential use is technically safe. VERIFY notes identify operational arrangements requiring editorial confirmation. Authors, reviewers and editors remain bound by the journal’s other ethics policies.

Scope and Definitions
AI Cannot Be an Author
Human Responsibility and Oversight
Disclosure of AI Use
Language Editing versus Content Generation
Verification of Facts and References
Fabricated Citations, Plagiarism and Attribution
AI in Research, Analysis, Data and Code
AI-Generated or AI-Assisted Images and Media
Privacy, Intellectual Property and Secure Use
Reviewers: Confidentiality and Responsible Use
Editors: Confidentiality and Responsible Use
Fairness, Bias and Accessibility
Detection and Evidence
Undisclosed or Inappropriate AI Use
Policy Updates and Contact

1. Scope and Definitions

Generative AI produces new output in response to instructions or data. AI-assisted technologies may translate, summarise, rewrite, classify, analyse, code, search, transcribe, enhance images or support other tasks. A product’s label does not determine the policy: the relevant questions are what information entered the system, what the system did, and how its output affected the work.

This policy covers authors and contributors, reviewers, editors and anyone acting for the journal. It does not treat every automated feature as equivalent. Ordinary deterministic functions such as spell-check, reference formatting or arithmetic may require less disclosure than a system that creates prose, interpretations, images, data or analytical results.

2. AI Cannot Be an Author

An AI system, chatbot or other automated tool must not be named or cited as an author or co-author. It cannot approve the submitted or published version, take public responsibility for the work, answer questions about integrity, consent to publication, disclose conflicts of interest, hold copyright or enter an authorship agreement.

Only humans who satisfy the journal’s Authorship and Contributorship Policy may be listed as authors. Use of AI does not create authorship credit for the tool, its developer or its operator. Human assistance that does not meet authorship criteria should be acknowledged where appropriate.

3. Human Responsibility and Oversight

The human authors are fully responsible for every part of the submitted work, including material drafted, translated, analysed or modified with AI. They must review the output, make the scholarly decisions, preserve the intended meaning and be able to explain and defend the methods and conclusions.

Responsibility cannot be transferred to an AI provider or excused by describing output as automatic. Authors must correct inaccurate, biased, plagiarised, unsafe or unsupported content before submission. They remain responsible for compliance with research ethics, consent, privacy, intellectual property, data availability and all declarations.

4. Disclosure of AI Use

AI use must be disclosed when it generated or materially transformed manuscript content, ideas, interpretations, translations, summaries, tables, figures, images, code, data processing, statistical analysis or research results. The disclosure should identify the tool and version where known, explain its purpose and the part of the work affected, and state that human authors reviewed and take responsibility for the output.

A disclosure belongs in the Methods when AI affected research methods, analysis, code, data or results; writing or presentation assistance may also be described in an Acknowledgements or other clearly visible statement. Relevant use should be reported to the journal at submission even if it is not ultimately described as a research method. Prompts or outputs needed to understand or reproduce a method should be preserved and supplied where lawful, ethical and safe.

VERIFY AI1 — Confirm the exact submission field or document in which authors should provide AI disclosures, who reviews them, and whether prompt/output supplements can be accepted and stored. No dedicated OJS disclosure form or retention workflow has been verified.

5. Language Editing versus Content Generation

Basic language assistance corrects spelling, grammar, punctuation or readability without introducing new claims, evidence, interpretations, citations or conclusions. Routine spell-checking and formatting that do not materially generate or transform scholarly content ordinarily do not require a detailed AI declaration. Authors nevertheless remain responsible for the final wording.

Content generation creates or substantially rewrites text, translates a meaningful part of a manuscript, proposes arguments or conclusions, summarises literature, generates references, or changes the intellectual substance. Such use requires disclosure and careful human verification. Labelling a generative function as “editing” does not remove the disclosure duty. If the boundary is uncertain, authors should disclose the use.

6. Verification of Facts and References

Authors must independently verify every factual statement, numerical value, quotation, attribution and reference affected by AI. Verification means consulting the actual source or underlying data, not asking the same or another AI system whether its output is correct. A fluent or confident answer is not evidence.

Each cited source must exist, be accurately described and support the claim for which it is cited. Authors should read the relevant source, check authorship, title, publication details, DOI or URL, and correct context. AI output may omit contrary evidence or amplify biases; authors must conduct an appropriate scholarly search and present a balanced account.

7. Fabricated Citations, Plagiarism and Attribution

Fabricated, non-existent, altered or misleading citations are prohibited, whether produced by a person or an AI tool. Submission of a reference that an author has not verified is not excused by AI assistance. Repeated or material false references may be handled under the journal’s Research and Publication Misconduct Policy.

AI-generated text or other output may reproduce protected or unattributed material. Authors must identify and cite original sources, obtain permissions where required, and ensure that plagiarism or inappropriate text recycling has not occurred. An AI system should not be cited as the source of a scholarly fact when a verifiable primary or authoritative source exists. Research specifically studying AI output should describe and preserve that output as research material in a transparent manner.

8. AI in Research, Analysis, Data and Code

When AI contributes to study design, participant interaction, data collection, coding, classification, modelling, statistical analysis or interpretation, authors must describe the role with enough detail for evaluation. They should document relevant inputs, software and version where known, settings or prompts material to the result, human validation, error checks and limitations, subject to ethical, legal and confidentiality constraints.

AI-generated or simulated information must never be presented as observed empirical data. Authors must distinguish original observations from synthetic, imputed or augmented data and justify their use. They must validate code and analytical output against the underlying data and suitable checks. AI use does not replace ethics review, informed consent, domain expertise or reproducibility obligations.

9. AI-Generated or AI-Assisted Images and Media

Images or media generated, reconstructed or materially altered by AI must not be presented as original experimental or observational evidence. Scientific images may be adjusted only in ways permitted by the journal’s ethics standards, without adding, removing, moving or concealing features that change interpretation. Enhancement that affects evidential meaning requires transparent description and access to appropriate originals.

A synthetic illustration may be considered only when it is clearly labelled, scientifically justified, legally usable and not liable to be mistaken for data. Authors should disclose the tool, purpose and material transformations and retain source files and generation records where appropriate. The journal may request originals or qualified assessment; detection output alone does not establish manipulation.

VERIFY AI2 — Confirm whether the journal accepts any AI-generated illustrative images, the required labelling/file formats, who conducts image assessment and how source files are retained. No specialist image-forensics service or approved image generator has been verified.

10. Privacy, Intellectual Property and Secure Use

Users must not enter personal, sensitive, confidential, copyrighted, embargoed or proprietary information into an AI system unless they have authority and an appropriate legal, ethical and security basis. They should examine relevant terms, data-use settings, retention, training, access and cross-border processing rather than assume that deletion or privacy is guaranteed.

Authors must not upload another person’s confidential material, private correspondence, peer-review report or unpublished work without authority. Consent for research participation does not automatically permit disclosure to an external AI provider. Removing names may be insufficient where the remaining information can identify a person or disclose protected intellectual property.

VERIFY AI3 — Confirm whether any institutionally controlled AI service is approved, the applicable data-processing and retention conditions, and who may authorise exceptional handling of protected information. No public or commercial AI service is approved by this page.

11. Reviewers: Confidentiality and Responsible Use

A reviewer must not upload a submitted manuscript, abstract, figure, supplementary file, author correspondence or review into an open, public, consumer or otherwise unauthorised AI service. A submitted manuscript is confidential. Uploading it may expose privileged content, personal data and intellectual property and may allow retention or reuse outside the review process.

Reviewers must perform the intellectual evaluation themselves and remain responsible for the accuracy, fairness, originality and tone of their report. They must not use AI to fabricate expertise, references or observations. Any AI-assisted use authorised by the journal must preserve confidentiality and be disclosed to the editor with the tool and purpose. If confidentiality cannot be assured, the tool must not be used.

12. Editors: Confidentiality and Responsible Use

Editors and persons acting for the journal must not upload confidential manuscripts, reviews, identities, editorial correspondence or unpublished decisions to an unauthorised AI system. They should not expose a manuscript to external model training, retention or human review without an appropriate basis and any permission required from the authors.

Editorial screening, summaries or suggested correspondence produced with an authorised AI tool cannot replace editorial judgement. The responsible editor must verify the output, identify bias or errors, protect identities and make the decision. AI must not make autonomous acceptance, rejection, reviewer-selection, ethics or sanctions decisions.

VERIFY AI4 — Confirm any authorised secure use by reviewers or editors, the approval authority, disclosure mechanism, confidentiality safeguards and recordkeeping. No journal-approved AI review or editorial system has been verified, so confidential material must not be uploaded to an external AI service.

13. Fairness, Bias and Accessibility

AI systems can reproduce bias, underrepresent languages or regions and give unequal results across populations. Authors, reviewers and editors must assess whether AI-assisted methods or text distort groups, erase uncertainty, reinforce stereotypes or disadvantage researchers because of language or access to paid tools.

A manuscript should be judged on its scholarly quality and compliance, not merely on a detector score or writing style. Reasonable language assistance should not be confused with authorship or misconduct. Questions about AI use should be specific, evidence-based and allow the author an opportunity to explain.

14. Detection and Evidence

AI-detection software may produce false positives and false negatives and must not be treated as conclusive proof that AI was used or that misconduct occurred. A concern should be assessed using the manuscript, references, data, disclosed methods, correspondence and other relevant evidence. Technical indicators must be interpreted by a competent person.

The journal may ask authors to explain their workflow, verify references, provide source material or clarify an AI disclosure. The request should identify the concern without presuming guilt. Refusal or inability to supply information may limit assessment, but the editorial decision must still be based on the available evidence and applicable policy.

15. Undisclosed or Inappropriate AI Use

Undisclosed material AI use, fabricated references, plagiarism, invented data, breach of confidentiality, deceptive image generation or misrepresentation of a method may require clarification and proportionate editorial action. The response depends on the conduct, evidence and effect on the reliability of the work; AI involvement does not create an automatic sanction.

For a submitted manuscript, action may include requesting disclosure or correction, further review or rejection where justified. For a published article, the journal may publish a correction, expression of concern or retraction under the Corrections, Retractions and Expressions of Concern Policy. Authors may use the Complaints and Appeals Policy for an applicable complaint or appeal.

VERIFY AI5 — Confirm final decision authority, the route when that person is conflicted, who updates notices and metadata, and the actual appeal route. No automatic sanction, fixed response period or guaranteed detection procedure is asserted.

16. Policy Updates and Contact

Because AI capabilities, risks and standards evolve, the journal may revise this policy prospectively. The version applicable to a submission should be interpreted with the instructions communicated at the relevant time. A later policy change should not by itself be used to label earlier conduct as misconduct.

Questions, disclosures or concerns may be sent to admin@srt-journal.uz. Include the manuscript reference or article DOI/title and only the information necessary to explain the issue. Current general contact information appears on the Contact page. Email delivery and monitoring are not guaranteed by this statement.

Read this policy with the Publication Ethics and Malpractice Statement, Authorship and Contributorship Policy, Research and Publication Misconduct Policy, Complaints and Appeals Policy and Corrections, Retractions and Expressions of Concern Policy.

Supporting Guidance

The journal-specific basis is its published ethics statement and related policies. The following guidance informs the framework; it does not prove that the journal has an unverified AI system, vendor arrangement or detection capability.

Contents