Responsible AI Use Policy

Preamble

JOTMI publishes research on technology, management and innovation at a time when research itself is undergoing an unprecedented technological transformation. The arrival of large language models (LLMs) and generative artificial intelligence tools has irreversibly altered the practices of scholarly writing, analysis, peer review and scientific communication. To ignore this shift would be a disservice to our readers; to react with blanket prohibitions would be naïve and counterproductive.

This policy takes a deliberate middle position: artificial intelligence is a legitimate, useful technology that, when properly employed, benefits the academic ecosystem. Its use is not, in itself, an editorial breach. What this journal requires —and verifies— is transparency, substantive human oversight, bibliographic integrity and empirical truthfulness. Authorship, together with the scientific and ethical responsibility for every manuscript, remain entirely within the human domain.

This document complements and is to be read jointly with the JOTMI Publication Integrity Protocol and Extended Retraction Policy, of which it constitutes the preventive and orientative counterpart.

1. Purpose and scope

This policy establishes the principles, definitions, obligations and recommendations governing the use of artificial intelligence at every stage of JOTMI's editorial cycle. It applies to authors, co-authors, peer reviewers, associate editors, members of the Editorial Board and the journal's technical staff. It covers manuscript preparation, submission, peer review, copy-editing, publication and post-publication oversight.

The policy applies regardless of the AI model used, whether open-source or proprietary, locally hosted or cloud-based, and regardless of its specific version. It also applies to tools that incorporate AI as a secondary layer (writing assistants, neural translators, reference managers with automated suggestions, peer-review assistants, image and figure generators, statistical-programming assistants, and so on).

2. Operational definitions

For the purposes of this policy:

Generative artificial intelligence: any computational system capable of producing original text, code, numerical data, images, audio or video from instructions (prompts) or input data. This includes, but is not limited to, general-purpose large language models (LLMs) and multimodal systems.

Workslop: textual or tabular content that is technically coherent but substantively empty, repetitive, evasive or disconnected from the actual contribution of the study, characteristic of automated generation without human curation.

Hallucination: a claim, reference, datum or citation produced by an AI system that has no correspondence with reality or with any verifiable source.

Ghost citation (V2): a bibliographic reference whose formal identifier (DOI, URL, ISBN) does exist but resolves to a document different from the one the author claims to be citing.

3. Guiding principles

Six irreducible principles articulate the entire policy:

Transparency. Any non-trivial use of AI in preparing a manuscript must be disclosed in a specific and verifiable manner. Deliberate omission is treated as a serious editorial breach.

Non-delegable human responsibility. Authorship is exclusively human. AI systems cannot be listed as authors or co-authors, cannot sign conflict-of-interest statements, cannot assume ethical commitments and therefore cannot be held accountable for content. Every signing party assumes full responsibility for every claim, figure, reference and visual in the manuscript, without exception and regardless of the tool used to produce them.

Truthfulness and empirical rigour. The use of AI does not, in any degree, attenuate the standard requirements of honesty, replicability and data traceability. Data fabricated by a model, non-existent references and hallucinated results are scientific fraud.

Substantive oversight. Any AI output incorporated into a manuscript must be read, understood, verified and, where appropriate, rewritten by the human author or authors. Nominal oversight —signing without reading— is equivalent to no oversight at all.

Equity and accessibility. JOTMI recognises that access to AI tools improves linguistic equity for authors whose first language is not English and for researchers with fewer institutional resources. This policy protects that access and rejects punitive interpretations that would, in practice, penalise the academic communities that benefit most from these tools.

Confidentiality. Any observation, finding, score or signal generated during the editorial process —including outputs from the automated audit infrastructure— is treated as strictly reserved information shared only among the editorial team, the authors and, where applicable, the assigned reviewers. The audit is a tool for editorial improvement that serves the author and the integrity of the scientific record; it is not a mechanism for public exposure. This confidentiality is unconditional: it does not depend on the author's compliance with this policy nor on the final decision regarding the manuscript.

4. Policy for authors

4.1 General stance

JOTMI does not reject or penalise manuscripts merely on the grounds of having been supported by artificial intelligence. The editorial question is not "did you use AI?", but "is what you assert true, original, rigorous and useful?"

4.2 Compatible uses without detailed disclosure

Purely instrumental uses with low influence on substantive content do not require detailed disclosure. These include: spelling and grammar correction, translation of the manuscript from the author's native language, lexical and stylistic suggestions, automated bibliographic formatting in reference managers, file-format conversion and audio transcription. A brief mention in the AI Use Declaration section (see 4.5) is sufficient when the author considers it appropriate.

4.3 Uses requiring explicit disclosure

The following uses must be specifically declared in the corresponding section of the manuscript (see 4.5):

• Substantive generation or rewriting of fragments of the manuscript text, beyond linguistic correction.
• Generation of abstracts by AI, in whole or in part.
• Assistance in formulating hypotheses, methodological design or interpretation of results.
• Generation of, or assistance with, statistical-analysis code, data-processing scripts or database queries.
• Primary-data analysis using AI tools (automated classification, AI-assisted qualitative coding, topic modelling, etc.).
• Bibliographic search delegated to generative AI systems, particularly when these have suggested references that the author then incorporated into the manuscript.
• Generation of figures, schematics, diagrams or images by generative AI.
• Substantial reformulation or restructuring of the manuscript.

The declaration must specify: which tool or model was used (with version, where available), in which sections, for what purpose and under what level of subsequent human oversight. This information is added to the editorial file; it does not, by itself, determine the outcome of the evaluation, except where it reveals breach of the prohibited uses listed in 4.4.

4.4 Prohibited uses

The following constitute serious editorial breach, subject to immediate rejection or, where applicable, subsequent retraction in accordance with the Extended Retraction Policy:

• Presenting as original any content generated entirely by AI without verification or substantive human rewriting.
• Attributing authorship —in whole or in part— to an AI system.
• Including hallucinated or non-existent bibliographic references, or references whose DOI resolves to a work different from the one cited (ghost citations V2).
• Fabricating numerical data, observations, interview transcripts, experimental results or any other empirical material by means of AI.
• Inserting hidden text, microtext, embedded comments or residual instructions (chat prompts) intended to manipulate the editorial process or subsequent indexing.
• Substituting Latin characters with homoglyphs from other alphabets (Cyrillic, Greek, etc.) in order to evade similarity-detection systems or manipulate searches.
• Artificially inflating the reference list with citations bearing no substantive relation to the text, including excessive self-citation or AI-suggested courtesy citations.
• Any practice consistent with paper mills, sale or purchase of authorship, hyper-prolific publishing without verifiable contribution under CRediT, or impersonation of academic identity.

4.5 Form and placement of the declaration

The AI use declaration must appear in a dedicated section titled "Declaration on the Use of Artificial Intelligence", placed immediately after the Acknowledgements and before the References. In manuscripts undergoing double-blind review, where Acknowledgements are removed in the anonymous version, the Declaration section is placed directly after the Conclusions. In manuscripts whose AI use does not exceed the compatible uses listed in 4.2, the following sentence suffices: "The authors did not use generative artificial intelligence tools in the substantive preparation of this manuscript."

4.6 Final responsibility

In application of the principle of non-delegable human responsibility (section 3), signature of the manuscript constitutes express declaration that each co-author has read the complete content and personally verified every cited bibliographic reference prior to submission. This verification is not transferred to, nor diminished by, the use of any automated tool.

5. Policy for peer reviewers

JOTMI peer reviewers perform an intellectual function whose delegation to AI systems, without disclosure or oversight, contravenes editorial trust and compromises the confidentiality of manuscripts.

Reviewers may use AI tools for ancillary tasks —verifying references, translating the manuscript into their working language, formatting comments— provided that: (a) they declare such use to the handling editor in the confidential-comments section of the report; (b) they do not upload the complete manuscript or substantial parts of it to AI services whose data-retention policies do not guarantee confidentiality; (c) the substantive opinion on contribution, methodology and relevance comes from the reviewer's human judgement, not from the model.

JOTMI considers, among other signals, the following as indicators of inappropriate delegation of review: response times systematically close to zero between assignment and report submission, reports exhibiting lexical patterns typical of AI-generated content, the absence of specific observations tied to the empirical content of the manuscript, and recurring patterns across reports produced by the same reviewer. Such signals will prompt private editorial enquiry to the reviewer, with no automatic sanction, but may result in removal from the panel if a systematic practice is confirmed.

JOTMI also reserves the right to incorporate internal AI-based editorial-assistance tools —including tools for reviewer suggestion, technical pre-evaluation of submissions and bibliographic verification— under the same principle of transparency that this policy demands of authors. Any substantive editorial use of AI will be reported in the institutional section of the journal's website.

6. Policy for editors and editorial staff

Associate editors, guest editors and the technical staff of JOTMI are bound by the same requirements of transparency and human oversight defined for authors and reviewers. Substantive editorial decisions —to accept, request revisions, reject or retract— are non-delegable human responsibilities. The audit tools described in section 7 generate technical evidence; they never, by themselves, issue editorial decisions.

The editorial team will keep an internal record of consultations to AI systems made in the course of the editorial process, particularly where these have influenced the final decision. This record will be available to the Editorial Board and, in the event of a documented author appeal, to the corresponding review procedure.

7. Automated editorial-audit infrastructure

In line with the standard practice of indexed academic journals, JOTMI uses commercial similarity- and plagiarism-detection tools recognised by the publishing industry (Crossref Similarity Check / iThenticate and equivalents). On that basis —and in a complementary, not substitutive manner— it operates its own audit infrastructure consisting of four interoperable modules covering dimensions that similarity detection alone does not address: authorship integrity, bibliographic truthfulness with DOI verification, patterns of automated text generation, and internal coherence between the data described, the methods declared and the results reported.

AAM (Authorship and Affiliation Module), currently at v2.2, performs verification of authorial identity, consistency of affiliations, ORCID traceability and detection of patterns consistent with paper mills. It includes a Document Integrity Scan phase (Phase 0.5) with safety toggles for the detection of homoglyphs, hidden text and structural file manipulation. It also applies the Source Recovery Protocol for cases of similarity-detection evasion.

BIM (Bibliographic Integrity Module), currently at v3.2, evaluates the integrity of the reference section through a composite index (BIX), the typological classification of each reference (10 categories), the detection of invalid DOIs, ghost citations, hijacked journals and hallucinated references, and verification of bibliographic currency relative to the field.

WAM (Workslop and AI-text Module), currently at v3.0 adapted to JOTMI, analyses the manuscript's lexical signature, characteristic patterns of automated generation without curation, the quality of academic English (AEQ) and the management-specific codes MG1 to MG3 proper to the journal's editorial line.

DSRM (Data and Statistical Reporting Module), currently at v2.0, operates on a verbatim extraction phase (Extraction Dossier) intended to prevent hallucinated evaluations, and verifies internal coherence between described data, declared methods and reported results.

These modules may be applied at any stage of the editorial process, including post-publication. Their results constitute technical evidence and are one —not the only— input to human editorial decision-making.

8. Consequences regime

Breaches of this policy are graded according to their severity and their wilful or negligent character, and are linked directly to the JOTMI Publication Integrity Protocol and Extended Retraction Policy.

Minor breaches —omission of disclosure where it was due, and template declarations that provide no real information— are typically resolved through a request for correction during the editorial process, without affecting the final decision.

Moderate breaches —invalid references not detected by the author, AI-generated sections without disclosure but verifiable as substantive, generic declarations that conceal intensive use— are resolved through a request for major revision or rejection, at the discretion of the handling editor.

Serious breaches —data fabrication, hallucinated bibliography, ghost citations V2, evasion techniques (homoglyphs, microtext, embedded prompts), attribution of authorship to AI, practices consistent with paper mills— activate the procedures of the Extended Retraction Policy, including definitive rejection at the pre-publication stage or formal retraction at the post-publication stage, in accordance with the Retraction Notice protocol and the permanent mark on the document.

9. Confidentiality of the editorial audit and official communications

JOTMI operates under a policy of strict confidentiality covering the entire editorial process. Manuscripts under evaluation, internal observations, scores produced by the audit infrastructure described in section 7, and correspondence between authors, reviewers and editors all constitute reserved information. This information is not disclosed to third parties, is not published, is not commented on social media, is not used for purposes of public exposure and does not constitute a basis for decisions outside the journal itself. The audit is an editorial tool serving the author and the scientific record; it is not a denunciation mechanism.

The formal consequences described in section 8 —when applicable— are executed only through the standard public mechanisms of the academic community, principally the Retraction Notice and the corresponding marks on the published document, with their traceability in bibliographic databases. JOTMI does not disseminate evaluation findings through informal channels, not even regarding rejected or withdrawn manuscripts, which retain the same level of information protection as an accepted manuscript.

Official communications. The only official communication of JOTMI with authors, reviewers and institutions takes place from the @jotmi.org domain. Any message received from other domains —including personal email accounts, visually similar but non-identical domains, social networks, instant messaging or alleged representatives contacting via informal channels— must not be regarded as official and must be reported to the editorial office through the official domain.

JOTMI does not request payments for processing manuscripts, does not require deposits, transfers or fees to expedite review, does not work through intermediaries, and does not condition any editorial decision on payment of any charge whatsoever. Any communication containing such a request must be treated as an attempted fraud and reported to the editorial office. This warning is particularly relevant for authors in regions where paper mills and extortion schemes posing as legitimate journals are widespread.

Calls for papers and special issues. JOTMI special issues originate exclusively from the initiative of the Editorial Board and are announced through the official site jotmi.org. JOTMI does not accept unsolicited proposals to coordinate special issues, nor external invitations to act as a guest editor received via personal accounts or any channel other than the institutional one. Authors who receive invitations to submit manuscripts to "JOTMI special issues" through unofficial channels must verify the call against the announcements published on the official site; any call not appearing there must be treated as spurious, with no need to reply to the sender. This directive responds to an international pattern of editorial fraud (the guest-edited special issue scam) extensively documented by COPE, Retraction Watch and the specialised science press since 2019.

10. Use of JOTMI content by external AI systems

JOTMI publishes its material under open licences that permit reading, distribution and reuse subject to attribution. The use of the journal's content for training artificial intelligence systems is a matter distinct from the free circulation of science, and JOTMI reserves the right to implement reasonable technical measures —including robots.txt configuration, server headers and other traffic-management tools— to regulate massive automated scraping of its site. These measures do not affect human accessibility of the content nor the legitimate rights of readers and authors.

11. Governance, validity and review

This policy enters into force on its date of publication and applies to all manuscripts submitted from that moment onward. Manuscripts already in the editorial process at the time of publication remain governed by the version in force at the time of their submission, except with regard to the consequences regime for serious breaches in section 8, which applies retroactively to the entire historical archive of the journal as part of post-publication audit.

The policy will be reviewed at least once per academic year by the Editorial Board, taking into account technological developments, the evolution of international standards (COPE, ICMJE, WAME) and the evidence gathered by JOTMI's own audit infrastructure. Substantive amendments will be versioned and archived; the version archive will remain publicly available.

Queries from authors, reviewers and institutions concerning the interpretation of this policy should be directed to the JOTMI editorial office. The journal commits to responding with consistent, public and traceable criteria.

Institutional document v1.0. Complementary to the JOTMI Publication Integrity Protocol and Extended Retraction Policy.