The use and limitations of artificial intelligence
The journal Anales de Teología recognizes that generative artificial intelligence systems can assist with certain writing and academic management tasks, but considers theological scholarship to involve formulating problems, constructing arguments, interpreting sources, and publicly defending a reasoned position. This policy therefore distinguishes between uses that address the form of a text written by its authors, which the journal permits with disclosure, and uses that replace the intellectual work underlying the contribution, which it does not permit. It applies to authors, reviewers, and editors, and complements the policies on Integrity, Originality, and Authorship, Editorial Process and Peer Review, and Editorial Ethics and Independence.
General principles
Artificial intelligence systems cannot be recognized as authors or coauthors because they cannot assume responsibility for a contribution, approve its publication, declare conflicts of interest, or manage rights. Authors bear full responsibility for the accuracy, originality, and legality of all submitted content, including any portions in which these tools have been used.
Editorial decisions, peer review, and the examination of possible breaches are human activities. No decision concerning a manuscript or a person will be based exclusively on the output of an automated system.
For the purposes of this policy, generative artificial intelligence means any system capable of producing, rephrasing, summarizing, or translating text, images, or other content in response to instructions. Spelling and grammar checkers that do not rephrase text, reference managers, and conventional search engines are not considered uses of generative artificial intelligence.
Use by authors
Uses permitted with disclosure
Authors may use generative artificial intelligence systems, with disclosure, to:
- improve the linguistic accuracy, clarity, or style of texts they have written themselves, without altering their argumentative content;
- translate their own texts, including the title, abstract, and keywords into English, provided they review the entire translation;
- locate potential bibliographic references or formulate search terms, provided they directly consult every work they use and do not replace reading it with system-generated summaries, descriptions, or interpretations;
- standardize reference formatting or adapt the manuscript to the journal’s editorial requirements;
- create charts or diagrams based on data, arguments, or conceptual relationships they have defined themselves, provided they verify that these accurately represent the content.
For translations published by the journal as contributions, the use of machine translation as an aid must be disclosed and does not exempt the translator from checking the entire translation against the original text. The journal may reject translations that do not demonstrate a thorough philological and conceptual review.
Prohibited uses
Generative artificial intelligence systems must not be used to:
- generate the thesis, arguments, interpretations of sources, or other content forming part of the intellectual work underlying the contribution;
- produce quotations, paraphrases, or references that have not been verified directly against the original source;
- present content generated or altered by these systems as an authentic source, testimony, or item of data;
- process research participants’ personal data or confidential information without the safeguards required by the Editorial Ethics and Independence policy.
The intellectual work underlying a contribution includes, among other elements, the formulation of the problem and thesis, the development of arguments, the analysis and interpretation of sources, the review of existing scholarship, the conclusions, the abstract, and substantive responses to reviewers’ comments. Translation and language editing of this content, when it has been written by the authors, are governed by the permitted uses.
Authors must verify every direct quotation and reference against the edition they state they have used. The inclusion of nonexistent references, quotations that do not match the source, or false attributions constitutes a breach of academic integrity, regardless of the tool that produced them.
Uses not addressed in this policy will be assessed according to its guiding distinction between form and intellectual work, and must be disclosed. When in doubt, authors may consult the Editorial Team before submission.
Disclosure
Any use of generative artificial intelligence must be disclosed in the declarations section required by the Integrity, Originality, and Authorship policy. The disclosure must identify the tool, its version and provider where available, the period of use, its purpose, and the sections or tasks in which it was used. It must also confirm that the authors reviewed the output and assume responsibility for the content. For example:
Use of generative artificial intelligence: The author used DeepL Pro, web version, in March 2026 to translate the title, abstract, and keywords into English. The author reviewed the translation and assumes full responsibility for the content of this work.
When none of these tools has been used, authors must use the statement specified in the Integrity, Originality, and Authorship policy. The journal may request additional information about a disclosed use, such as the instructions provided or intermediate versions of the text, when necessary to assess the integrity of the work.
Artificial intelligence as an object of study
When a contribution analyzes texts or outputs produced by artificial intelligence systems as an object of theological research, it must identify the system, its version, the dates of the queries, and the instructions used, and make the analyzed outputs available within the text or as supplementary material. Such outputs may be cited only in these circumstances and must never be treated as authoritative sources on the subject under discussion.
Peer review
For reasons of confidentiality and responsibility for scholarly judgment, reviewers must not enter, upload, or share manuscripts or associated materials, in whole or in part, in generative artificial intelligence systems, regardless of the purpose. They must not use these systems to assess the manuscript or draft or rephrase their reports, which must express their own judgment. When submitting their reports, reviewers must confirm that they have complied with these conditions.
Editorial Team
The same restrictions apply to the Editorial Team and anyone involved in handling manuscripts. The journal may use tools that do not generate content, such as similarity detection and reference verification systems, and may employ generative artificial intelligence systems for administrative tasks that do not involve unpublished manuscripts or personal data, under the supervision and responsibility of the Editorial Team.
Undisclosed or prohibited uses
Given their margin of error, the results of tools that estimate the probability that a text was generated by artificial intelligence do not in themselves constitute proof of misuse. The journal may consider them as an indication alongside verifiable evidence such as nonexistent references, quotations that do not match the source, or false attributions.
When undisclosed or prohibited use is suspected, the journal will request an explanation from the authors and proceed in accordance with the Complaints, Appeals, Corrections, and Retractions policy. Before publication, it may request clarifications or revisions, or reject the manuscript; after publication, it may publish a correction, an expression of concern, or a retraction, depending on the seriousness of the case. Undisclosed substantive use is considered a breach of academic integrity.
Updates
The journal will review this policy when the available tools, COPE guidance, or the practices of the theological community change, and will keep the date of each version publicly available.
This policy draws on COPE’s position on authorship and artificial intelligence tools.
