Allegations of Research Misconduct Policy

Purpose & Scope

ARCITECH: Journal of Computer Science and Artificial Intelligence is committed to evaluating credible concerns about the integrity of submitted and published research, peer review, authorship, research data, and the publication process. This policy applies to manuscripts, published articles, supplementary materials, datasets, code, editorial communications and relevant research records.

The purpose of this procedure is to protect readers and the scholarly record, ensure fair handling of concerns, and enable proportionate editorial action. A report is an allegation, not a finding that any individual has committed misconduct.

Meaning of Research and Publication Misconduct

Research misconduct may involve fabrication, falsification, plagiarism, serious research-ethics breaches, or other deliberate or reckless misrepresentation of scholarly evidence. Publication misconduct may include deceptive authorship practices, manipulated peer review, fabricated citations, undisclosed redundant publication, or other interference with an honest editorial process.

Not every research error, disagreement, failed replication or methodological limitation constitutes misconduct. Honest error and questionable research practice require assessment according to their effect on the reliability of the record, and do not automatically establish an individual’s intent.

Examples of Reportable Concerns

The following examples are not exhaustive. Each report is assessed in its factual and disciplinary context.

• Fabrication or falsification of datasets, sample counts, model outputs, experimental measurements, statistical analyses, citations or reported results.
• Plagiarism, deceptive text recycling, unacknowledged appropriation of code, figures, images, tables, methods or data, and material redundant publication.
• Manipulation of scientific figures, image/video/audio evidence, screenshots or synthetic materials presented misleadingly as authentic observations.
• Peer-review fraud, fabricated reviewer accounts, coercive citations, paper-mill activity, fabricated references, or coordinated publication-process manipulation.
• Guest, gift, ghost, purchased or fictitious authorship; material undisclosed conflicts of interest, funding or research-ethics approvals.
• Unlawful or unethical acquisition or disclosure of personal information, restricted cybersecurity logs, system credentials or other sensitive data.

Who Can Report and Where

Authors, readers, reviewers, editors, research institutions, funders or other persons may submit a documented concern. ARCITECH accepts good-faith reports relating to work under review or already published, regardless of whether the reporter is associated with the journal.

Send reports to the editorial office at arcitech.journal@iaincurup.ac.id. Suggested subject: Research Misconduct Allegation — [Manuscript ID / Article DOI]. If the concern involves the Editor-in-Chief or another person with decision-making authority, explicitly request confidential referral to an unconflicted publisher or institutional representative.

Information and Supporting Evidence

A useful report identifies the specific article or manuscript and separates observations from interpretations. Unsupported accusations and threats are not treated as substantiated findings.

• Manuscript title and submission ID, or published article DOI, URL and bibliographic information.
• Precisely described issue: relevant text, figure, table, dataset, code file, reference, experiment or review event, including location and date where known.
• Supporting material such as legitimate comparison sources, reproducibility logs, verifiable screenshots, version history, permission records or correspondence.
• Explanation of why the concern may materially affect reliability, research ethics, authorship or editorial independence.
• Reporter contact details where available, and any request for discretion, with reasons for anonymity or restricted evidence.

Initial Screening and Triage

An appropriately designated editor or editorial representative records the report and evaluates whether it falls within journal responsibilities, contains actionable particulars, calls for expert review, or presents an immediate risk to readers, individuals or information systems.

The editor may seek clarification, preserve relevant records, and determine whether a manuscript should temporarily stop progressing while a serious concern is examined. A similarity score or AI-text detector score is a screening signal, not sufficient proof of misconduct. The journal may close a report that is demonstrably irrelevant or lacks actionable evidence, with a brief explanation where practicable.

Independent Handling and Recusal

Anyone named in an allegation or with a material financial, personal, academic or institutional conflict must not control the assessment, appoint assessors, review evidence selectively or decide the outcome. A qualified and unconflicted editor shall handle the case whenever practicable.

When an allegation involves the Editor-in-Chief, an editorial-board member, reviewer, or a manuscript authored by a journal officer, oversight should be referred to an independent person designated by the publisher/institution. The journal records the recusal and replacement decision without disclosing confidential details publicly.

Evidence-Based Assessment Workflow

The journal applies the following sequence, adapted to the type and seriousness of the case. Stages may overlap when necessary to protect sensitive information or the integrity of a published record.

Response Targets and Communication

Consistent with ARCITECH’s Complaints & Appeals Policy, the editorial office aims to acknowledge an identifiable, sufficiently described report within five working days and communicate an initial substantive outcome or progress update within 30 calendar days. These are service targets—not guaranteed deadlines for completing a complex misconduct investigation.

The parties should receive reasonable updates when further expert, institutional or legal review is required, subject to confidentiality restrictions. The journal may act more quickly when immediate reader, privacy or cybersecurity risk exists.

Verification in Computer Science and AI

For AI/ML, data science, computer vision, software engineering and cybersecurity manuscripts, the editorial assessment may require context-specific technical records that can help establish whether reported results are reliable.

• Dataset provenance, dataset licensing, sample or class counts, train/validation/test splits, preprocessing order and evidence concerning data leakage.
• Code revision identifiers, package/library versions, model architectures, random seeds, computational logs, evaluation scripts and model checkpoints when legally and practically available.
• Clarification of baselines, hyperparameters, ablation studies, uncertainty estimates, benchmark configurations and reproducibility limitations.
• Original unaltered images or signals, manipulation descriptions and records distinguishing synthetic or generative content from actual observation.
• Authorization and safe handling requirements for network traces, exploit research, behavioral logs, personal information, credentials, unpatched vulnerabilities and other dual-use artifacts.

Whistleblowers and Anonymous Reports

Reports submitted anonymously, through a third party or originating in a public scientific forum can be considered when they contain enough specific evidence to check. A person should not be dismissed solely because they prefer anonymity, but lack of contact information may limit clarification and feedback.

ARCITECH seeks to protect a good-faith reporter’s identity to the extent practicable and lawful, while preserving a fair response opportunity for people named in the report. Absolute anonymity or immunity from lawful disclosure cannot be guaranteed. Knowingly fabricated allegations and harassment are not acceptable.

Author Response and Procedural Fairness

People whose work or conduct is questioned should normally receive a clear, neutral statement of the relevant concern and a reasonable opportunity to provide explanations and original supporting material. The journal shall avoid language that presumes wrongdoing before the evidence is assessed.

Independent technical reviewers may be consulted when specialized expertise is necessary. Review findings must be assessed critically, not treated as automatically determinative. Disagreement about a scientific interpretation alone is not a finding of misconduct.

Cooperation with Institutions and Authorities

The journal may ask an author’s institution, employer, research-integrity office, funding body or other competent authority to investigate matters beyond its editorial remit, such as primary research misconduct, employment discipline, disputed laboratory records or serious legal violations.

ARCITECH remains responsible for deciding what editorial action is required to protect its published record. Institutional findings may inform these decisions, but delays or limits in institutional investigations do not necessarily prevent proportionate public notices when credible concerns require readers to be informed.

Outcomes Before Publication

Depending on the assessment, possible outcomes include clarification, additional evidence, corrected analysis, revised ethics declarations, renewed independent review, rejection, or referral to an institution or competent authority.

An unresolved serious allegation may justify pausing consideration while relevant material is examined. Rejection is not itself proof of misconduct. Editors document the rationale and communicate decisions through authorized OJS correspondence.

Outcomes After Publication

For a published article, the reliability and significance of the concern determine the appropriate editorial response. The journal may take no public action where claims are unsupported; publish a correction or clarification for a material yet correctable error; publish an Expression of Concern where serious unresolved concerns merit notice; or retract a work whose main findings or publication integrity cannot be trusted.

Retraction and other public notices must be accurate, dated, traceable and clearly linked to the relevant article in accordance with ARCITECH’s Corrections, Retractions & Expressions of Concern Policy. A retraction serves to correct the scholarly record, not to determine individual punishment; the original record should not be silently erased.

Confidentiality, Records and Data Protection

ARCITECH seeks to keep submissions, unpublished evidence, peer-review identities, correspondence and personal data confidential on a need-to-know basis, subject to lawful obligations and fair inquiry. Records may include the original report, screening notes, conflict declarations, evidence requests, responses, independent assessments and reasoned decisions.

Sensitive records should be accessed only by authorized personnel. Retention, disclosure and deletion must follow applicable institutional policy and the journal’s Privacy Statement & Personal Data Protection Policy. The journal does not promise a specific technical safeguard or fixed retention period unless verified through its actual institutional arrangements.

Distinguishing Related Procedures and Appeals

A research-misconduct allegation concerns the integrity of evidence or publishing practices. A scientific disagreement can instead be addressed through Post-Publication Discussion; a procedural challenge to the journal through Complaints & Appeals; and a necessary change to an already published record through the Corrections, Retractions & Expressions of Concern Policy.

Individuals affected by an editorial outcome may seek reconsideration through the published appeals procedure by presenting new material evidence or identifying a substantive procedural error. Where practicable an unconflicted decision-maker should consider the appeal. Filing an appeal does not guarantee acceptance or automatically suspend a warranted public notice.

Contact, Related Policies and Standards

Reports and supporting inquiries: arcitech.journal@iaincurup.ac.id. Cite the manuscript ID or article DOI and mark messages involving journal leadership for independent escalation.

This policy must be read together with ARCITECH’s Publication Ethics, Peer Review, Conflicts of Interest, Complaints & Appeals, Corrections & Retractions, Research Data, and Privacy policies. Reference to international ethical standards does not imply accreditation, certification or membership by those organizations.