Similarity, Plagiarism and Investigation Policy

Similarity, Plagiarism and Investigation Policy

The International Journal of Artificial Intelligence, Educational Technologies and Teacher Education – IJAIET adopts a permanent policy for the prevention, identification, and investigation of textual similarity, plagiarism, self-plagiarism, redundant publication, and other forms of improper appropriation of scientific content, with the aim of preserving originality, research integrity, copyright, and the reliability of the scholarly record.

All submitted manuscripts may be screened using specialized similarity-detection tools, including systems such as iThenticate/Crossref Similarity Check, in addition to qualitative assessment by the Editorial Team.

IJAIET accepts, as an editorial screening parameter, a global similarity index of up to 30%, provided that the identified matches are legitimate, properly cited, and do not constitute plagiarism, relevant self-plagiarism, substantial reproduction of previous works, or any other violation of research integrity.

The journal emphasizes that 30% similarity does not mean 30% plagiarism. The percentage generated by software is only a technical indicator and must be editorially interpreted in light of the origin, location, extent, nature, and context of the matches.

This policy is guided by the good practices of the Committee on Publication Ethics – COPE, the FAPESP Code of Good Scientific Practice, the CNPq Research Integrity Guidelines, the recommendations of the Council of Science Editors – CSE, and the principles of ethics and research integrity adopted by IJAIET.

1. Concept of Similarity

Textual similarity refers to the identification, through computational tools or editorial assessment, of passages in a manuscript that match content available in articles, books, theses, dissertations, websites, repositories, institutional documents, or other sources.

The presence of similarity does not automatically constitute plagiarism.

Matches may legitimately arise from:

  • bibliographic references;
  • names of institutions;
  • technical terminology;
  • names of methods;
  • research instruments;
  • legislation;
  • normative documents;
  • properly identified direct quotations;
  • expressions established within a scientific field;
  • standardized descriptions of procedures;
  • materials previously made available as preprints, provided they are properly disclosed;
  • methodological content whose repetition is technically unavoidable and properly referenced.

For this reason, IJAIET does not make editorial decisions exclusively on the basis of the overall similarity percentage.

2. Editorial Similarity Limit

IJAIET establishes the following general parameter:

Permitted global similarity: up to 30%

A manuscript may proceed in the editorial workflow when it presents up to 30% total similarity, provided that the qualitative assessment confirms:

  1. absence of plagiarism;
  2. appropriate attribution of sources;
  3. absence of substantial reproduction from a single source;
  4. absence of relevant self-plagiarism;
  5. absence of duplicate or redundant publication;
  6. originality of the scientific contribution;
  7. independence of the argumentation, analysis, results, and conclusions.

Even below 30%, the manuscript may be returned, rejected, or referred for investigation when matches incompatible with good scientific practice are identified.

Likewise, exceptional cases above this percentage may be editorially assessed when the similarity results, for example, from standardized instruments, legislation, technical documents, protocols, theses deposited in repositories, disclosed preprints, or other justifiable elements. Such assessment does not imply acceptance.

Therefore:

The 30% threshold functions as an editorial screening parameter and not as authorization to reproduce content.

3. Similarity with a Single Source

The Editorial Team will pay particular attention to the concentration of similarity in a single source.

A manuscript may have an overall similarity index below 30% and still contain inappropriate reproduction of substantial portions of a specific work.

The following will be assessed:

  • extent of matching passages;
  • presence or absence of citation;
  • nature of the source;
  • location of the matches;
  • degree of verbatim reproduction;
  • authors’ own intellectual contribution.

Large blocks reproduced from a single source may constitute an integrity problem regardless of the overall percentage reported by the software.

4. Concept of Plagiarism

For the purposes of IJAIET policy, plagiarism is the use or presentation of words, ideas, arguments, data, images, methods, results, or other intellectual contributions of third parties without proper acknowledgment of authorship or source, leading readers to understand such content as the authors’ own work.

Plagiarism may be complete or partial and may involve:

  • verbatim reproduction without citation;
  • overly close paraphrasing without attribution;
  • appropriation of ideas;
  • improper use of data;
  • reproduction of tables, images, or graphs;
  • translation of another person’s content without citation;
  • use of third-party methods or intellectual structures without attribution;
  • appropriation of manuscripts, reviews, or confidential documents.

Plagiarism will be identified through qualitative analysis of the evidence, rather than solely by the similarity score generated by software.

5. Self-Plagiarism and Text Recycling

Self-plagiarism, or inappropriate text recycling, occurs when authors reuse substantial portions of their own previously published work without adequate transparency, citation, or justification.

IJAIET recognizes that limited reuse may occur, particularly in methodological descriptions. However, authors must:

  • cite their previous works;
  • indicate when the same data, samples, or projects have been used;
  • avoid extensive reproduction of previous text;
  • clearly demonstrate the original contribution of the new manuscript;
  • disclose related articles already published or under evaluation.

Substantial reuse of the introduction, results, discussion, or conclusions from previous work may constitute redundant publication or inappropriate self-plagiarism.

6. Duplicate and Redundant Publication

Submission or publication of work that substantially reproduces previously published content without scientific justification and editorial transparency is prohibited.

IJAIET may consider publication redundant when there is significant overlap in:

  • objectives;
  • hypotheses;
  • samples;
  • data;
  • results;
  • analyses;
  • tables;
  • figures;
  • discussion;
  • conclusions.

Manuscripts derived from the same research project, thesis, dissertation, or dataset must present an autonomous and clearly distinct scientific contribution.

Authors must disclose the existence of related works.

7. Translation Plagiarism

Translation into another language does not transform previously published content into original content.

It may constitute plagiarism or redundant publication to translate texts, substantial passages, or entire articles of third parties—or even one’s own work—and present them as unpublished content without proper identification of the original source.

IJAIET adopts trilingual publication in Portuguese, English, and Spanish. Official language versions produced or authorized by the journal are not considered duplicate publications, as they constitute versions of the same scientific publication linked to the same editorial record.

8. Similarity in Texts Produced or Modified by Artificial Intelligence

The use of Generative Artificial Intelligence does not remove authors’ responsibility for originality and integrity.

AI systems must not be used to:

  • conceal plagiarism;
  • rewrite copied text for the purpose of evading detection;
  • fabricate references;
  • create nonexistent citations;
  • generate paraphrases of third-party works without attribution;
  • appropriate copyrighted material;
  • fabricate data or results.

Content produced with AI support must fully comply with IJAIET’s Artificial Intelligence Use Policy.

Authors remain responsible for verifying originality, references, accuracy, copyright, and any potential matches generated by the system used.

9. Similarity Analysis Procedure

The process may include:

Stage 1 – Screening

The manuscript is submitted to a similarity-detection tool.

Stage 2 – Quantitative Assessment

The Editorial Team reviews the overall percentage and distribution of matches.

Stage 3 – Qualitative Assessment

The following are examined:

  • sources;
  • extent;
  • location;
  • citations;
  • concentration of matches;
  • methodological or conceptual nature of the passages;
  • possibility of self-plagiarism;
  • possible redundancy.

Stage 4 – Editorial Classification

The manuscript may be:

  • cleared to proceed;
  • returned for correction;
  • referred for clarification;
  • referred for an integrity investigation;
  • rejected.

10. Procedure When Similarity Exceeds 30%

When the report indicates similarity above 30%, the manuscript will undergo specific editorial assessment.

IJAIET may:

  1. request a reduction in similarity;
  2. require correction of citations;
  3. request genuinely original rewriting;
  4. request clarification regarding identified sources;
  5. request information on authors’ previous works;
  6. assess possible self-plagiarism or redundancy;
  7. reject the manuscript when originality is compromised;
  8. initiate an investigation into possible misconduct where there are indications of plagiarism, fraud, or deliberate concealment.

Superficial changes of wording made solely to artificially reduce the similarity percentage will not be regarded as adequate correction.

11. Procedure When Similarity Is Below 30%

A similarity index below 30% does not guarantee automatic approval.

The manuscript will still undergo qualitative assessment and may be rejected even at a low percentage when there is, for example:

  • extensive reproduction from a single source;
  • appropriation of a central idea;
  • copying of results;
  • reproduction of tables or images;
  • undisclosed translation;
  • plagiarism of an innovative methodology;
  • substantial self-plagiarism;
  • deliberate textual manipulation intended to avoid detection.

12. Investigation of Suspected Plagiarism

When significant concerns arise, IJAIET will follow an investigation procedure based on the principles of impartiality, confidentiality, documentation, right to respond, and proportionality, in accordance with COPE and its guidance on suspected plagiarism.

The procedure may include:

1. Identification of the Issue

Recording the sources and passages considered potentially problematic.

2. Preservation of Evidence

The similarity report, manuscript, related sources, communications, and other relevant records will be preserved.

3. Preliminary Assessment

The Editorial Team will determine whether there is sufficient basis to continue the investigation.

4. Notification of Authors

Authors will be informed of the concerns and given the opportunity to provide explanations and supporting documents.

5. Assessment of Responses

Authors’ explanations will be compared with the available evidence.

6. Editorial or Technical Opinion

When necessary, independent specialists in research integrity, methodology, copyright, or the relevant scientific field may be consulted.

7. Decision

IJAIET will determine the appropriate and proportionate editorial measure.

13. Possible Decisions Before Publication

When the issue is identified before publication, possible actions include:

  1. clearing the manuscript to proceed;
  2. requesting adjustments;
  3. mandatory correction of citations;
  4. rewriting specific sections;
  5. a new similarity analysis;
  6. editorial warning;
  7. temporary suspension of evaluation;
  8. rejection for lack of originality;
  9. rejection for plagiarism;
  10. referral of the matter to the authors’ institution when the seriousness of the case justifies it.

14. Plagiarism Identified After Publication

If plagiarism or related problems are identified after publication, IJAIET will investigate the case under its Policy on Research Misconduct, Investigation, Corrections, and Retractions.

Depending on severity, the journal may publish:

  • Correction;
  • Erratum;
  • Expression of Concern;
  • Retraction.

COPE recommends retraction where plagiarism is sufficiently substantial to compromise the legitimacy and integrity of the publication.

As a general rule, a retracted article will remain available and clearly marked as RETRACTED, preserving the traceability of the scholarly record.

15. Possible Sanctions

Where a deliberate and serious violation is established, IJAIET may, on a proportionate basis:

  • issue a warning;
  • reject the manuscript;
  • terminate editorial processing;
  • require correction;
  • publish a retraction;
  • temporarily restrict new submissions;
  • notify the authors’ institution;
  • notify the funding agency;
  • notify a Research Ethics Committee or other competent authority;
  • apply other editorial measures provided for under its integrity policies.

IJAIET does not impose institutional sanctions on behalf of FAPESP, CNPq, universities, or other organizations. Where necessary, the matter will be referred to the competent institution for assessment within its own jurisdiction.

16. Authors’ Responsibility

By submitting a manuscript, authors declare that:

  • the content is original;
  • all sources have been properly acknowledged;
  • no improper appropriation of content has occurred;
  • their own previous works have been cited when necessary;
  • the manuscript does not constitute duplicate publication;
  • the same manuscript is not simultaneously under consideration by another journal;
  • any preprints have been disclosed;
  • references have been verified;
  • AI use has been disclosed in accordance with journal policy.

Responsibility for manuscript originality remains entirely with the authors, regardless of whether IJAIET conducts a similarity analysis.

17. Responsibilities of Reviewers and Editors

Reviewers who identify possible plagiarism, redundant publication, or appropriation of ideas must report the concern confidentially to the Editorial Team, identifying relevant sources whenever possible.

Reviewers must not conduct public investigations or contact authors directly regarding the concern.

Editors must:

  • assess evidence impartially;
  • preserve confidentiality;
  • avoid conflicts of interest;
  • provide authors with an opportunity to respond;
  • distinguish legitimate similarity from plagiarism;
  • apply proportionate measures;
  • preserve the integrity of the scholarly record.

18. No Universal Percentage for Plagiarism

IJAIET emphasizes that there is no universal mathematical percentage that automatically determines the presence or absence of plagiarism.

Therefore, the journal’s editorial threshold of up to 30% similarity is an internal screening and editorial-management parameter, not a definition of plagiarism established by COPE, CNPq, FAPESP, or CSE.

The final assessment will always be qualitative, contextual, and evidence-based.

It is therefore incorrect to state:

“Up to 30% plagiarism is permitted.”

That statement does not represent IJAIET policy.

The correct formulation is:

IJAIET accepts a global similarity index of up to 30% as an editorial screening parameter, provided that the matches are legitimate, properly attributed, and do not constitute plagiarism, relevant self-plagiarism, redundant publication, or any other violation of research integrity.

19. Institutional References

COPE – Committee on Publication Ethics

COPE – Plagiarism in a Submitted Manuscript

COPE – Plagiarism in a Published Article

COPE – Retraction Guidelines

FAPESP – Good Scientific Practices

CNPq – Research Integrity Guidelines

CNPq – Commission on Integrity in Scientific Activity

Council of Science Editors – Recommendations for Promoting Integrity in Scientific Journal Publications

20. Summary Policy Statement

All manuscripts submitted to IJAIET may undergo similarity analysis. The journal adopts a global similarity index of up to 30% as an editorial screening parameter, subject to mandatory qualitative assessment of the identified matches. A similarity percentage does not automatically correspond to plagiarism. Even manuscripts below 30% may be rejected where plagiarism, substantial self-plagiarism, redundant publication, improper appropriation of ideas, data, images, or any other violation of research integrity is identified. Manuscripts above 30% will undergo specific editorial assessment and may be returned for correction, investigated, or rejected according to the nature of the matches.

Copyright Statement

© IJAIET – International Journal of Artificial Intelligence, Educational Technologies and Teacher Education / Editora Ibero-Americana de Educação. All rights reserved regarding the wording, organization, systematization, and institutional presentation of these editorial policies, without prejudice to third-party rights and content derived from duly identified external sources.

The full or substantial reproduction, adaptation, or institutional use of these texts by third parties requires prior authorization, except within the limits legally permitted for quotation, academic reference, and other uses authorized by applicable law.

The Editors