Similarity and AI
Similarity, Academic Integrity and AI Declaration. Version 1.0 · 2026. Complements the Author Guidelines and the AI Use Policy.
Academic integrity is a founding principle of PanAmerican Journal of Digital Humanities & AI. This page sets out the similarity checks, the criteria used to assess them, the consequences of plagiarism and the specific protocol for declaring the use of artificial intelligence. Similarity and AI are treated as two sides of one commitment: that every work published in this journal is what it claims to be.
Part I. Similarity control and plagiarism prevention
1. Similarity check at submission. All manuscripts are checked with Turnitin or equivalent software as part of the preliminary editorial review, before they go to peer review. Authors must attach the similarity report generated by the software when they submit. A submission without this report is returned to the author without review.
2. How the report must be configured. The report must be generated with these exclusions:
| Element to exclude | Reason | Required |
|---|---|---|
| Reference list | References repeat titles and data from other works and create structural similarity unrelated to the manuscript’s original content | Yes |
| Properly marked quotations | Quotations in quotation marks or in block form, with correct attribution, are a legitimate use of other authors’ text under APA 7th edition | Yes |
| Continuous strings of up to 20 words | Short strings of common or technical words that appear in many texts of the field are not plagiarism | Yes |
| Keywords and field terms | The specialized language of digital humanities and AI is shared by the community | Recommended |
| Title, authorship and affiliations | Identification data are not evaluable intellectual content | Recommended |
Authors may apply further exclusions if they are justified and stated in the Similarity and AI Form that accompanies the submission.
3. Thresholds and decisions.
| Similarity index | Status | What happens |
|---|---|---|
| 0–20% | Acceptable | The manuscript continues to the preliminary editorial review and then to peer review |
| 21–30% | Review zone | The Editorial Team examines the detailed report to determine whether the similarity comes from properly cited sources or from unattributed reproduction. The author may be asked for clarification |
| Over 30% | Rejection | Immediate rejection without peer review, except in very exceptional circumstances that justify a detailed manual review at the discretion of the Editor-in-Chief |
The index is an indicator, not an automatic verdict. An 18% index made up mostly of fragments from one source without attribution is more serious than a 25% index spread across several properly cited sources. The Editorial Team always evaluates the full report and not only the overall percentage.
4. Conduct that is subject to sanction.
| Conduct | Description |
|---|---|
| Direct plagiarism | Literal reproduction of another author’s text without quotation marks or attribution |
| Paraphrase plagiarism | Rewording another author’s text without attribution, keeping the same structure or ideas without citing the source |
| Self-plagiarism | Substantial reproduction of one’s own previously published work without declaring it as a source (also called duplicate publication) |
| Data fabrication | Presenting data, results or references that do not exist or were not generated by the process described |
| Falsification | Changing real data, results or processes to alter the conclusions of the research |
| Duplicate submission | Submitting the same manuscript to two or more publications at the same time |
5. Consequences and procedure. At any sign of plagiarism or other conduct that compromises academic integrity, the journal acts following the procedures of the Committee on Publication Ethics (COPE):
| When it is detected | Consequence |
|---|---|
| During the preliminary review | Immediate rejection. Written notice to the author with the reasons. Possible exclusion from future submissions for the period the Editorial Team decides |
| During peer review | Rejection of the manuscript. Notice to the author and, where appropriate, to the reviewers. Investigation following COPE |
| After acceptance | Withdrawal of the article before publication. Notice to the author. Possible notice to the affiliated institution |
| After publication | Public retraction of the article, with an explanatory note visible on the article page and in its permanent record. Notice to the indexing databases that include it. Communication to the affiliated institution following COPE |
Part II. Declaration of artificial intelligence use
This part complements the journal’s AI Use Policy. It sets the operating requirements for declaring AI use in the context of the similarity and integrity review.
Guiding principle. In this journal, declaring the use of artificial intelligence is an act of methodological rigor, not a confession. Transparency about AI use does not penalize the author: it accredits the work. Articles that describe precisely how AI was used in the research are valued for that contribution.
1. The three levels and their relation to the similarity check.
| Level of AI use | Expected effect on similarity | How the report is read |
|---|---|---|
| Level 1: Assistive | Minimal or none. AI improves form, not content | The report is assessed with the standard criteria. No additional justification is needed |
| Level 2: Methodological | Variable. Similarity may appear in methodology sections if AI processed existing literature | The author states in the form which sections were processed with AI. The Editorial Team assesses similarity in those sections with a differentiated criterion |
| Level 3: Collaborative | Possibly higher in synthesis or theoretical framework sections if AI contributed substantially | The author must describe in detail how the generated content was checked and edited. The Editorial Team assesses the quality of that supervision, not only the similarity percentage |
2. What must be declared and what need not.
- No declaration required (Level 1): grammar and spelling improvement with tools such as Grammarly or DeepL; assisted translation of drafts into academic English, especially for non-native authors; the word processor’s automatic spell checker; automatic formatting of citations and references. Declaring such use is never prohibited.
- Declaration in the methodology section (Level 2): thematic or discourse analysis of interviews or documents; inductive or deductive coding of qualitative data; automatic transcription of interviews or audio; systematic literature review with tools such as Elicit, Consensus or ResearchRabbit; assisted bibliometric analysis; generation of data visualizations; statistical modeling or generation of analysis code. The declaration states the tool, its version, the specific task and the validation method.
- Dedicated “Artificial Intelligence Use Statement” section (Level 3): literature synthesis with assisted generation of an argument or theoretical framework; structuring sections of the manuscript with substantial participation of AI; predictive modeling with assisted interpretation; generation of complex analysis code with verification of the results. The section describes the tools, the tasks, the researcher’s critical supervision and, optionally, a methodological reflection on the use of AI.
3. Absolute limits linked to similarity and AI. The following are incompatible with the journal’s integrity policy, whatever level of AI use is declared:
- presenting as original a text generated entirely by AI without substantial intellectual intervention by the researcher;
- using AI to reword or paraphrase other authors’ text in order to lower the similarity index artificially;
- generating bibliographic references with AI without checking that they exist (language models can produce citations that do not exist);
- not declaring Level 2 or Level 3 use, which violates the editorial policy and is handled following COPE;
- using non-anonymized personal data of human participants as input for AI tools.
Part III. Similarity and AI Form
A completed Similarity and AI Form must be attached as a separate document to every submission, whatever the level of AI use. A submission without it is returned without review. This single form replaces separate similarity and AI declarations. It has four parts: (A) the manuscript data; (B) the similarity report, with the index obtained, the software used and the exclusions applied; (C) the declaration of AI use, with the level and, for Levels 2 and 3, the tools, versions, stages, tasks and validation methods; and (D) the author’s ethical commitments and signature. The form can be downloaded from the journal’s Submissions page.
