1. General Principles and Policy Scope

This policy addresses submissions, revisions, editorial screening, peer review, publication preparation, and post-publication activities. It applies equally to all fields expressly recognized in ABIM's Aims & Scope. AI-related requirements do not impose an additional marketing or industrial-marketing variable where the manuscript already falls within a recognized field.

1. Human accountability. Authors, reviewers, and editors remain responsible for their contributions, assessments, recommendations, and decisions.
2. Transparency. Substantive generative-AI assistance must be accurately disclosed in accordance with the applicable journal requirements.
3. Scholarly integrity. Assistance must not produce fabricated evidence, misleading interpretations, unsupported claims, false citations, or plagiarized content.
4. Research ethics. AI use must respect informed consent, institutional approval, participant rights, organizational confidentiality, and applicable data-protection requirements.
5. Independent peer review. Automated evaluations must not substitute for assessment by qualified independent human reviewers.
6. Editorial independence. Scientific evaluation and publication decisions remain under authorized human editorial responsibility.

Relevant technologies include generative language models, AI writing and coding assistants, automated content-generation systems, and tools that materially transform research-related content. Routine conventional software functions are distinguished from substantive generative-AI assistance.

2. AI Usage Policy for Authors

2.1. Authorship and Intellectual Responsibility

AI systems cannot qualify as authors or co-authors because they cannot approve a final manuscript, accept scholarly accountability, disclose conflicts of interest, or respond to integrity concerns. Human authors must make genuine intellectual contributions, approve the work, and remain responsible for its originality, theoretical foundations, methods, analysis, interpretation, references, and conclusions.

AI assistance does not diminish authorship responsibilities or replace accurate CRediT contributorship reporting.

2.2. Permitted Uses of AI

Subject to appropriate human verification, disclosure, ethics, permissions, and confidentiality safeguards, legitimate assistance may include:

• Language editing: Improving spelling, grammar, clarity, and readability without changing scientific meaning.
• Translation assistance: Translating an author-prepared draft while independently checking terminology, conceptual equivalence, methodological accuracy, and argument.
• Technical or programming support: Assisting with analytical code, debugging, calculations, or visualization, with independent testing against the actual data and methods.
• Research-related methods: Using AI as an instrument, analytical system, computational method, or research object where justified and transparently documented.
• Explanatory graphics: Preparing accurate conceptual diagrams or clearly identified illustrations without misrepresenting them as observed or experimental evidence.

2.3. Prohibited Uses of AI

• Substituting automated output for genuine scholarly argument, critical literature synthesis, interpretation, or independent intellectual contribution.
• Fabricating respondents, consumer behavior, survey responses, interview transcripts, organizational records, financial data, market observations, or statistical results.
• Presenting synthetic cases, simulated observations, generated images, or artificial research materials as authentic empirical evidence.
• Manipulating tables, figures, numerical values, models, or outputs so that they misrepresent the underlying research.
• Including fictitious or unverified publications, quotations, bibliographic information, or DOI records.
• Uploading identifiable participant data, customer records, confidential business information, or restricted research materials to unauthorized or insecure services.
• Misrepresenting substantial AI-generated content as an independently developed human contribution without adequate scholarly work, verification, and disclosure.

2.4. Verification of Evidence and References

Authors must verify factual claims, theoretical interpretations, calculations, analytical outputs, quotations, citations, and references introduced or modified through AI assistance. Bibliographic suggestions are not verified sources. Original publications, relevance, author details, metadata, and DOI information must be checked before citation.

References must follow APA 7th edition and correspond with in-text citations in both directions. Analytical outputs must be validated against actual data, appropriate procedures, and defensible methodological standards. Human review must address the substance of the output, rather than merely its fluency or appearance.

2.5. Research Papers, Literature Review Papers and Conceptual Papers

These requirements apply to all three ABIM manuscript types: Research Paper, Literature Review Paper, and Conceptual Paper. AI techniques, analytical tools, and bibliometric methods do not create additional manuscript categories.

Research Papers must accurately describe the system's methodological role, validation, limitations, and relevant reproducibility information. Literature Review Papers require independent checking of publication identification, selection, assessment, and synthesis; generated summaries do not replace critical reading of original sources.

Conceptual Papers require genuine scholarly reasoning, accurate theoretical attribution, coherent development of concepts or relationships, and independent evaluation of proposed contributions. AI-generated propositions or frameworks must not be treated as established knowledge without critical justification.

3. AI Usage Policy for Editors

3.1. Human-Led Editorial Decision Making

Authorized human editors retain responsibility for acceptance, rejection, revision requirements, and scientific evaluation. Decisions must address the manuscript's recognized field, contribution, integrity, appropriate methods or conceptual reasoning, and independent review evidence. Automated systems must not independently determine scholarly merit.

3.2. Permitted Administrative Assistance

Where authorization, confidentiality, and oversight are assured, automated tools may support administrative checks without replacing academic judgment. Potential functions include:

• Checking formatting, required files, declarations, and submission completeness.
• Supporting secure similarity screening or preliminary reference checks, followed by human examination.
• Identifying potential reviewer expertise from appropriately accessible scholarly information.
• Improving routine editorial correspondence without exposing protected submission information.
• Organizing administrative checklists and workflows under appropriate access controls.

These examples describe possible controlled uses; they do not claim that ABIM has deployed or approved any particular AI product or processing service.

3.3. Editorial Restrictions

• Do not upload confidential manuscripts, unpublished data, participant information, reports, or correspondence to public or insecure AI systems.
• Do not delegate substantive scientific evaluation or publication decisions to generative AI.
• Do not treat an automated AI-detection probability as a substantiated finding of misconduct.
• Do not treat generated scientific comments, references, criticisms, or recommendations as verified evidence without examination.
• Do not use external services in ways that compromise confidentiality, intellectual property, or peer-review integrity.

3.4. Editorial Oversight

Potential concerns identified by automated screening require evidence-based human assessment and appropriate clarification. Screening indicators must be interpreted in context. Manuscripts selected for external assessment normally proceed through double-anonymous review by at least two independent qualified human reviewers.

4. AI Usage Policy for Peer Reviewers

4.1. Independent Expert Judgment

Reviewers are selected for subject knowledge, methodological competence, and independent scholarly assessment. Each report must represent the reviewer's own examination of the manuscript. Responsibility for evaluating contribution, methods, evidence, conceptual reasoning, and conclusions cannot be transferred to AI.

4.2. Permitted Limited Assistance

Limited assistance with the language or clarity of independently written comments may be considered only where journal authorization and confidentiality requirements are satisfied. Removing author names alone does not necessarily make unpublished manuscript information safe to disclose. Any proposed external processing of confidential material requires explicit editorial authorization and adequate safeguards.

4.3. Prohibited Reviewer Practices

• Uploading manuscripts, abstracts, tables, figures, extracts, interview material, datasets, or confidential correspondence to public or insecure services.
• Requesting a complete AI-generated report as a substitute for independent scholarly review.
• Delegating methodological assessment, interpretation, or publication recommendations without independent evaluation.
• Submitting criticisms based on invented references, unsupported arguments, or content the reviewer has not examined.
• Using privileged manuscript information for personal, professional, competitive, commercial, or unrelated research purposes.

4.4. Reporting Suspected AI Misuse

Reviewers should report specific evidence of fabricated references, invented data, manipulated findings, or other material concerns confidentially to the responsible editor. They must not independently accuse authors, contact their institutions, or regard an automated detector output as definitive proof of misconduct.

5. Confidentiality, Data Security and Research Ethics

ABIM requires protection of participants, unpublished manuscripts, restricted datasets, institutional records, and confidential editorial materials. Relevant contexts include consumer research, customer databases, employee records, commercial transactions, financial information, organizational strategy, and other sensitive business or social research.

• Participant protection: Protect identifiable and sensitive information in accordance with applicable consent and ethical requirements.
• Organizational confidentiality: Do not disclose restricted customer, employee, financial, contractual, or institutional records without appropriate authority.
• External systems: Do not upload confidential data to public or insecure services. Prior editorial authorization and adequate safeguards are required for proposed external processing of confidential editorial content.
• Research ethics: Respect ethics approval, informed consent, data-use permissions, institutional authorization, and third-party rights.
• Incident reporting: Report suspected unauthorized disclosure or misuse promptly through appropriate editorial or institutional channels.

Where synthetic responses, simulated markets, generated scenarios, or artificial observations are legitimate research objects, identify their artificial origin, describe the procedures, and avoid presenting them as genuine participant or organizational data.

6. Declaration and Disclosure of AI Use

6.1. When Disclosure Is Required

Under ABIM's supplied policy, substantive generative-AI assistance beyond routine spelling, grammar, formatting, or straightforward language correction must be disclosed. Material assistance affecting analysis, research procedures, conceptual development, translation, or manuscript content should be described accurately.

PERLU VERIFIKASI: The treatment of specific borderline uses, including extensive language restructuring and incidental coding assistance, should be confirmed against the journal's formally adopted implementation guidance. Uncertainty should be disclosed to the Editorial Office rather than concealed.

6.2. Where the Declaration Must Be Placed

Submission placement: Include the declaration in the separate Title Page & Author Information Form. Keep author-identifying information out of the Blinded Manuscript.

When AI materially contributes to research design, data collection, coding, analysis, experiments, or review methodology, include sufficient non-identifying methodological information in the manuscript for scientific assessment. This does not introduce a third mandatory submission file.

6.3. Declaration Components

• Tool identification: Identify the tool or service; include the developer and relevant model, version, or access date where known and important for reproducibility.
• Purpose and scope: Explain the tasks, manuscript sections, or research procedures affected.
• Human verification: Describe how relevant content, code, calculations, references, or analytical outputs were checked and corrected.
• Human responsibility: Confirm responsibility for originality, integrity, accuracy, interpretation, and conclusions.

The tool and purpose are established disclosure requirements in ABIM's supplied source. The additional detail above provides implementation guidance and must accurately reflect the work actually performed.

6.4. Standard Declaration Formats

The following formats describe the information needed for common uses. They are complete drafting guidance, not declarations of AI use by any particular author. Authors must provide their own accurate account.

Disclosure does not authorize fabricated evidence, insecure disclosure, plagiarism, or other prohibited conduct. Do not state that verification occurred unless it actually did.

7. Policy Violations, Investigation and Sanctions

7.1. Fair Investigation and Due Process

AI-related concerns must be assessed objectively and proportionately through ABIM's research-integrity procedures. Automated classifications and unsupported allegations do not establish misconduct. The journal may examine manuscript materials, sources, records, methods, declarations, and explanations, with appropriate confidentiality and conflict management. Affected parties should have a reasonable opportunity to respond.

7.2. Possible Actions Concerning Authors

Permitted assistance with an incomplete declaration may require clarification or correction. Verified misuse affecting originality, references, methods, or reliability may require substantive revision or rejection. Serious post-publication concerns may require a correction, Expression of Concern, or retraction according to the evidence.

PERLU VERIFIKASI: A dedicated AI-specific submission restriction or sanction schedule has not been established by the supplied ABIM source. No automatic ban, fixed penalty, or retrospective sanction is created by this page. Any additional action requires an appropriate documented decision under applicable governance.

7.3. Possible Actions Concerning Reviewers

A report affected by unsupported automated evaluation or a confidentiality breach requires independent editorial assessment of its reliability and the need for replacement review.

PERLU VERIFIKASI: Formal warning, termination of an assignment, or removal from the active reviewer pool are proposed personnel responses from the master framework. Their authority and application must be confirmed under ABIM governance rather than presumed to be an existing automatic sanction.

7.4. Possible Actions Concerning Editors

An editor must not determine the outcome of an investigation involving their own conduct or a material conflict. Independent handling is required.

PERLU VERIFIKASI: Reassignment, withdrawal of handling authority, formal corrective action, or review of an appointment must follow confirmed ABIM governance arrangements and the established evidence. This page does not independently confer disciplinary authority.

7.5. Post-Publication Actions

Where concerns arise after publication, ABIM may seek author explanations, examine supporting records, consult independent specialists, and take an appropriate documented action. Corrections, Expressions of Concern, and retractions must be identifiable and linked to the original article. The Version of Record must not be silently altered to conceal material concerns.

8. Responsibilities and Implementation

This page complements ABIM's authorship, peer-review, plagiarism, confidentiality, research-ethics, complaints, and post-publication provisions. The separate page's adoption and effective date remain PERLU VERIFIKASI. The continuous-publication date must not be reused as its effective date without evidence.

Historical manuscripts, editorial records, valid publications, and bibliographic information are not automatically invalidated by subsequent policy changes. Concerns about earlier work must be considered on their evidence under the applicable procedures.