Similarity and AI

1. Foundation and Editorial Philosophy

Artificial intelligence is today a reality in the research process. This is not an emerging trend: it is an established fact that the international scientific community recognizes, uses, and, in its great majority, practices in silence. A recent study in the Proceedings of the National Academy of Sciences (2026) analyzed millions of articles published after 2023 and found that while more than 50% of journals have some type of AI policy, fewer than 0.1% of articles declare having used it. The gap between actual practice and official declaration is, in that context, a problem of scientific integrity — not of technology.

This journal has chosen a different path: rather than prohibiting what cannot be verified, we have decided to build a framework that makes the declared use of AI an act of methodological rigor. Rather than treating AI as inherently suspect, we treat it for what it is: a tool that, well used, extends the capabilities of the human researcher without replacing their thinking, judgment, or responsibility.

There is also a dimension of justice that this policy explicitly recognizes. Scientific research has historically been a privilege of those with time, resources, and infrastructure. A researcher at a well-funded university in the Global North can spend months reviewing literature, coding qualitative data, or processing large-scale surveys. An equally brilliant researcher in Latin America, with limited database access, multiple teaching loads, and no support staff, cannot always compete on the same terms. Artificial intelligence does not eliminate that gap, but it reduces it significantly. Ignoring that potential for reasons of institutional tradition is not prudence: it is conservatism disguised as rigor.

2. The Five Principles of This Policy

Principle I — AI is a Methodological Tool

This journal treats the use of artificial intelligence tools with the same rigor and normalcy as the use of any other research instrument: statistical software, computer-assisted content analysis, automatic transcription, reference managers. None of those tools is questioned for its mere existence. AI will not be either.  The question this journal asks is not «whether AI was used» but «how it was used, with what criteria, and what was the researcher's critical contribution to its results».

Principle II — The Researcher is the Sole Intellectual Author and Sole Responsible Party

No artificial intelligence tool may be named as author, co-author, or official collaborator of an article published in this journal. This position is non-negotiable and aligns with the international consensus of COPE, ICMJE, APA, and Nature Portfolio.  Intellectual and ethical responsibility rests entirely with the human author or authors. This includes responsibility for any error, bias, or inaccuracy present in the work, regardless of whether it was generated or processed with AI assistance.  There are dimensions of the research process that belong exclusively to the human researcher: the formulation of the research question, theoretical positioning, interpretation of the meaning of findings, conclusions, and the ethical, political, and social implications of the work.

Principle III — Declaring AI Use is an Act of Rigor, Not Confession

This journal deliberately inverts the dominant narrative. Declaring the use of AI tools in a research process is not an admission of weakness, a confession of shortcut, or a signal of lower intellectual quality. It is, within the framework of this publication, a demonstration of methodological transparency, scientific confidence, and commitment to the reproducibility of the research process.  Articles that precisely and in detail describe how AI was used in their process will be recognized for that rigor during editorial evaluation.

Principle IV — Epistemic Equity as an Editorial Value

Artificial intelligence has a redistributive potential that this journal explicitly recognizes. It allows a researcher with limited resources to systematically review thousands of articles, analyze extensive interview corpora, process data from multiple countries, and communicate findings in high-quality academic English, without any of those capabilities implying delegation of scientific thinking.  This journal recognizes that potential as an academic good. Not as an unfair advantage but as a partial correction of the structural inequalities that have limited, for decades, the participation of Latin American researchers in the highest-impact international scientific conversations.

Principle V — Reflection on AI Use is New Knowledge

This journal is, in itself, a publication dedicated to digital humanities and artificial intelligence. From that perspective, methodological reflection on how AI was used in a research process is not a bureaucratic appendix: it is an original contribution to the field.  Authors are invited to consider, when relevant, including a reflection on what the use of AI made visible that would otherwise have been invisible; on the biases it may have introduced; on how it affected the speed, scope, or depth of the analysis. That reflection is not mandatory, but it is welcome and editorially valued.

3. The Three-Tier Model

The use of artificial intelligence in research is not homogeneous. There is a substantial difference between using AI to improve the grammar of a text and using it to code two hundred in-depth interviews. This policy establishes three levels of use, each with differentiated declaration requirements. The central logic is simple: the greater the AI's participation in the intellectual process, the greater the precision with which it must be described.

4. Declaration Protocol by Research Type

4.1 Qualitative Research: In-Depth Interviews

The analysis of in-depth interviews with AI assistance is one of the most powerful and documented uses in recent methodological literature. A researcher who has used AI for thematic coding, discourse analysis, or category extraction from interview transcripts must declare in the Methodology section:

• The AI tool used and its version.
• Whether the transcripts were anonymized before being entered into the tool.
• The coding criteria or analytical frameworks the researcher defined in advance.
• How the validation of the generated categories was carried out (review by the lead researcher, triangulation, second coder, etc.).
• Whether the process was iterative: how many review cycles were conducted.

4.2 Systematic Literature Review

The use of AI tools to support literature review — including initial article screening, extraction of key data, and narrative synthesis — is a methodological use that this journal fully recognizes. The declaration must specify which tools were used, at what phase of the process, and how the researcher validated the results. It is recommended, when possible, to include the list of search terms used as input for the AI tool.

4.3 Quantitative Analysis and Data Modeling

The use of AI to generate statistical analysis code (Python, R), interpret model results, or visualize data constitutes a Level 2 methodological use. The researcher must declare which tool was used, for what task, and how they verified the validity of the code and results produced. The researcher is responsible for the model design, interpretation of results, and derived conclusions.

4.4 Academic Writing in a Second Language

This journal has an explicit linguistic equity policy. The use of AI tools to improve the academic English of authors whose first language is Spanish, Portuguese, or French is a Level 1 use and requires no declaration. The thinking, argument, and research belong to the author. The communication tool is an accessory — not a substitute for intellect.  This position is anchored in a principle of equity: no valuable research should be rejected or degraded because its author did not master a language that is not theirs from childhood.

5. Absolute Limitations

The following provisions are mandatory and admit no alternative interpretation:

6. AI Use Declaration Form

Every manuscript submitted to PanAmerican DH&AI must include the following completed form as a separate document at the time of submission. If AI was not used at any level, the author must explicitly indicate this.

7. Context: Why This Framework Matters for the Americas

Latin America faces a paradox: it possesses extraordinary cultural, social, and human wealth, with urgent problems requiring rigorous research and evidence-based solutions. At the same time, regional scientific production has historically been underrepresented in the highest-impact international publications.

The reasons are multiple and well-known: lack of funding, teaching overloads for researchers, limited access to international databases, language barriers, and the gap between the time rigorous research requires and the resources available to conduct it. A researcher working under those limitations is not less capable — they are more valuable, because their questions arise from contexts that the Global North does not know from the inside.

Artificial intelligence can act as a capability multiplier. It does not eliminate the need for critical thinking, theoretical knowledge, cultural sensitivity, or ethical commitment. But it can compress weeks of routine work — systematic bibliographic review, interview transcription, initial data coding — into hours. And that frees the researcher to do what no algorithm can: think deeply, interpret with context, and propose solutions with wisdom.