Peer Review Policy & Editorial Decision

ARCITECH: Journal of Computer Science and Artificial Intelligence evaluates original research manuscripts through a transparent, confidential, and independent editorial process. Editorial assessment is based on scientific merit, relevance to the journal’s aims and scope, originality, methodological rigor, reliability of findings, and compliance with publication ethics.

This policy applies to research articles, review articles presenting scholarly research, and other submissions identified by the journal as peer reviewed. Editorials, notices, and administrative material that are not externally peer reviewed must be clearly identified as such. Any exception to ordinary review arrangements must be transparently communicated.

ARCITECH uses a double-blind peer-review model: reviewers are not informed of the authors’ identities, and authors are not informed of reviewers’ identities during the editorial process. The handling editor has access to author information as necessary to administer the review and manage conflicts of interest.

Every research article sent for peer review is assessed by at least two independent, appropriately qualified subject specialists who are not members of the journal’s editorial team for that submission. Reviewers must be selected on expertise, independence, and capacity to provide a timely, reasoned evaluation.

The journal takes reasonable steps to anonymize manuscripts and supporting files, while recognizing that publicly available preprints, repositories, or prior related research may sometimes make identities inferable. Reviewers must not attempt to identify authors or contact them directly.

Reviews are treated as confidential editorial records; reviewer names and unpublished review reports are not ordinarily published alongside articles.

Upon submission through OJS, the editorial team conducts an initial screening before external review. The screening considers:

• Alignment with the journal’s stated focus and scope, article type, and submission requirements.
• Completeness of manuscript files, metadata, anonymization, required declarations, and references.
• Originality, clarity of the research question, substantive scholarly contribution, and baseline methodological adequacy.
• Potential ethical concerns involving authorship, conflicts of interest, personal data, research permissions, or duplicate submission.
• Similarity screening using Turnitin, with results interpreted by an editor in context rather than treated as stand-alone proof of plagiarism.

ARCITECH’s existing guidance specifies a 20% similarity benchmark. The editor examines matched passages, quotations, references, technical terminology, and contextual evidence; similarity percentages do not automatically establish or exclude misconduct.

Manuscripts that fail the initial requirements may be returned for technical correction or declined without external peer review. A desk rejection is an editorial decision, not a reviewer recommendation.

The handling editor appoints reviewers whose substantive expertise matches the manuscript. Reviewer identity, affiliation, publication expertise and independence should be reasonably verified before invitation. Reviewer suggestions from authors may be considered but never guarantee appointment.

Reviewers must disclose relationships that could compromise or reasonably appear to compromise their judgment, including current shared employment, recent coauthorship, direct supervision, close personal relationships, financial interests, active research competition, or ongoing collaboration. The handling editor makes an appropriate recusal or reassignment decision.

• A reviewer must decline an assignment outside their expertise or capacity.
• Reviewers must not delegate an assignment to another person without prior editorial authorization.
• Editors must avoid fabricated reviewer identities, unauthorized changes to reviewer email addresses, and manipulated reviewer reports.
• Only properly completed, independent reports count toward the minimum of two peer reviews.

Reviewers are expected to give specific, constructive, evidence-based comments addressing, as applicable:

• Contribution and novelty: clear research gap, significance, and appropriate comparison with related work.
• Technical validity: sound algorithms, software and system design, statistical analysis, evaluation protocol, and justified assumptions.
• Experimental rigor: appropriate datasets, baselines, controls, metrics, uncertainty assessment, and justified interpretation of results.
• Replicability: sufficient information on data preparation, software versions, hyperparameters, hardware, procedures, and limitations.
• Responsible research: data permission and privacy, intellectual-property rights, security impacts, research ethics and disclosure of limitations.
• Scholarly presentation: coherent structure, accurate figures/tables, relevant references, and well-supported conclusions.

ARCITECH states a four-week review period for manuscripts that pass desk screening. This is an operational target, not a guaranteed deadline: reviewer availability, methodological complexity, and additional review rounds may affect the actual duration.

Editors monitor reviewer invitations, acceptance, and submitted reports in OJS. When significant delays arise, the editorial office should inform authors, give a reasonable status update, and where necessary invite an alternative qualified reviewer.

The time required for desk screening, author revision, additional review rounds, and production is separate from the initial review target. Authors can request a status update through the journal’s official communication channel.

Reviewers advise the editor. Their recommendations do not constitute acceptance or rejection by the journal. The editor evaluates the evidence across reports, the manuscript, ethical considerations, and author responses before recording the decision.

An editor may issue a different decision from the reviewers’ recommendations when there are documented scholarly or ethical grounds. Such decisions should be reasoned, consistent, and recorded within the OJS workflow.

Following the currently published ARCITECH process, authors are generally given 14 calendar days for minor revisions and 28 calendar days for major revisions, counted from the editorial revision request unless the decision letter states an adjusted date.

Authors should upload a revised manuscript and a point-by-point response identifying the action taken for each editor or reviewer comment, including the location of changes. Where a recommendation is not followed, authors must explain their scientific rationale respectfully.

An author who needs more time should request an extension before the deadline. The editor may approve a reasonable extension or close the submission after appropriate communication if the revision is not received. Revised manuscripts may be assessed by the editor, the original reviewers, or additional independent reviewers, depending on the issues raised.

Following screening and review, the designated handling editor may communicate one of the following editorial outcomes:

The Editor-in-Chief or a formally delegated, conflict-free handling editor is responsible for the journal’s final scientific decision. For normal submissions, the Editor-in-Chief retains oversight of editorial standards. Acceptance is not automatic following favorable reviewer reports, and post-acceptance production does not waive responsibilities for research integrity.

When reviewer recommendations conflict substantially, the handling editor must assess the substance and validity of each report rather than merely averaging votes or following a majority. The editor may request further clarification, invite a third independent reviewer, request targeted revision, or make a reasoned decision when the evidence is sufficient.

Additional reviewers do not replace the requirement to document independent scientific review. Disagreements about scientific methods should be resolved by considering validity, reproducibility, evidence and relevance—not author status or financial considerations.

Manuscripts authored or coauthored by editors, editorial board members, reviewers, or institutional affiliates are subject to the same desk screening, two-reviewer standard, methodological criteria, and editorial thresholds as other submissions.

An editor who is an author, collaborator, or otherwise substantially conflicted must recuse themselves from reviewer selection, review management and the editorial decision on that manuscript. A suitably qualified independent editor must manage the case. If the Editor-in-Chief is conflicted, a conflict-free designated editor or qualified independent guest editor must oversee the decision.

The publisher or journal management must preserve a verifiable record of the delegation and conflict handling. Affected editors must not influence the decision through their managerial position. ARCITECH should monitor the share of published work authored by internal editors, board members and reviewers in line with DOAJ’s endogeny criteria.

Reviewers and editors should confidentially report specific concerns about plagiarism, duplicate submission or publication, fabricated or falsified data, manipulated citations, figure tampering, stolen text or code, manipulated peer review, or other unethical practices.

Editors evaluate reasonable evidence, request clarification from the author where appropriate, and may seek specialist advice. An unresolved integrity concern may justify pausing review, requesting evidence, rejecting a manuscript, or taking other actions under the journal’s Publication Ethics & Malpractice Statement. Accusations must not be treated as proven without fair examination, and confidentiality should be protected.

Peer reviewers and editors must exercise independent human judgment. They must not upload unpublished manuscripts, confidential review reports, private correspondence, or non-public datasets to external generative-AI services without documented authorization and adequate confidentiality safeguards.

Generative AI must not be used to produce a substantive scientific review or replace an editor’s decision. Limited assistance with the reviewer’s own non-confidential wording may be permitted when confidentiality is preserved and any relevant use is disclosed to the editor.

Automated similarity or integrity alerts require human evaluation and cannot serve as the sole evidence for a misconduct finding.

Editorial decisions should be communicated through the journal’s authorized OJS or official email channels. Where appropriate, decision letters provide the outcome, reasons for significant concerns, anonymized reviewer feedback, revision requirements, and the next editorial step.

Reviewer identities must remain confidential under the double-blind model. Editors should ensure feedback forwarded to authors is professional and does not expose confidential reviewer information or contain defamatory or discriminatory remarks. Submitted, accepted, and published dates should be accurately recorded and displayed on published research articles where the OJS workflow supports them.

Authors who believe that a decision involved a demonstrable factual error, material procedural irregularity, unmanaged conflict of interest, or other substantive concern may submit a reasoned appeal to the official ARCITECH editorial address. The request should identify the manuscript and supply specific supporting evidence.

Disagreement alone with an unfavorable scientific opinion does not constitute grounds for reversal. Where practicable, an appeal is reviewed by a qualified editor or publisher representative who was not responsible for the contested decision. A substantiated appeal may result in clarification, independent reassessment, or reopening of the editorial process; it does not guarantee acceptance.

Complaints about the Editor-in-Chief should be referred for conflict-free consideration at the publisher level. Correspondence and outcomes should be documented confidentially.

If ARCITECH publishes special issues, proceedings-derived manuscripts, or guest-edited collections, research articles must meet the same scope, originality, ethical, and minimum two-independent-reviewer requirements as regular issues. Guest editors must be subject to editorial oversight and conflict-of-interest controls.

Any research article authored by a guest editor must be handled independently. Editorial notes, announcements and other material not sent for independent peer review must be labeled to avoid misleading readers about the review received.

Acceptance moves the manuscript to copyediting, proofreading and production. Authors may be asked to check the final proof for factual and typographical accuracy. The journal then schedules the article for its relevant issue or continuous-publication release under the applicable publication policy.

OJS editorial records should document reviewer invitations, competing-interest declarations, review reports, revision files, decisions, correspondence and delegations. Review records are confidential and accessible only to authorized journal personnel for editorial and compliance purposes.

APC invoices and payment arrangements are handled independently of reviewer recommendations and scientific decisions. See the journal’s Article Processing Charges (APC) page.

For peer-review questions, reasonable status inquiries, formal appeals or concerns about editorial conduct, please contact:

Policy development references:

• DOAJ — Guide to Applying (version 2.9).
• Principles of Transparency and Best Practice in Scholarly Publishing.
• COPE — Ethical Guidelines for Peer Reviewers.
• PKP / OJS — Review Stage and Editorial Actions.

Reference to these organizations and resources does not imply certification, endorsement, or DOAJ indexing of ARCITECH.