New UOC policy for a transparent, responsible and ethical use of artificial intelligence in its journals

Generative artificial intelligence has taken the world of academia and publishing by storm. In just a few years, text generation, summarizing, translation and image creation tools have become part of many researchers and authors' daily work. While they create new opportunities, these technologies also raise questions about authorship, content quality, transparency and scientific integrity.

To address these new challenges, the Universitat Oberta de Catalunya (UOC) has approved the Policy on the use of artificial intelligence (AI) for writing and reviewing articles published in the UOC's scientific and general-interest journals, which establishes the criteria to be followed by both authors and reviewers when using AI in the publication process. The policy, which was approved by the Research and Knowledge Transfer Committee and the UOC's Executive Board in 2025, applies to both the university's scientific journals – Artnodes, IDP, DIGITHUM and Dictatorships and Democracies – and its general-interest journals ComeIn, Oikonomics and Mosaic.

The aim is to provide a clear framework to ensure that authors and reviewers are aware of their responsibilities when using AI, as well as upholding transparency, ethics and the integrity of research.

A regulatory framework that does not seek to ban the use of AI

The advent of generative artificial intelligence has opened up new opportunities for research and academic publishing, but it has also raised questions about authorship, content reliability and the responsibility of those who put their name to the resulting publications. It is against this backdrop that the UOC has chosen to establish a framework regulating the use of these tools rather than banning them outright.

According to Ayllen Gil, a member of the Faculty of Law and Political Science, researcher with the DITD (Law, Internet and Digital Transformation) research group and director of the journal IDP. Journal of Internet, Law and Politics, "the main challenge lies not only in establishing authorship but in ensuring the integrity of published content in the face of unreliable AI detection tools and the existence of AI tools capable of creating anything from coherent writing to plausible but fabricated experimental data." She explained that the aim of the policy is precisely to "set a clear boundary to ensure that technical assistance does not replace intellectual rigour".

This is also the philosophy behind the document: the aim is not to preclude the use of AI but to ensure that it is used transparently and under human responsibility. Mariló Martín, director of the Academic Publications unit at the UOC, believes that a blanket ban would only lead to greater opacity, whereas requiring authors to disclose when and how they have used these tools will foster trust in scientific publishing. In her opinion, AI can be a valid support tool provided it remains under human control.

Transparency from the outset

The policy is based on seven principles: transparency, integrity and ethics, equity and non-discrimination, human supervision, human compatibility, quality and legality. They all support the core concept that, although it can be used as part of the editorial process, AI can never replace human critical judgement or responsibility.

One of the main changes introduced by the policy is that any author who uses AI will have to expressly disclose this in the manuscript, specifying which tool they used and for which tasks and purposes, such as to assist with writing, convert information between different formats or structure their arguments.

The policy also states that AI tools must be cited in accordance with each journal's referencing conventions, just like any other source used in the research process.

Responsibility remains in human hands

One of the core pillars of the policy is that responsibility for content should never lie with technology. AI tools cannot be recognized as authors or co-authors of articles, and using them does not release authors from their responsibility for the accuracy, quality and integrity of their published work.

The policy emphasizes that any AI-generated text must be critically reviewed before publication, as AI tools may produce incorrect, incomplete or biased information (commonly known as "hallucinations"), as well as discriminatory outputs or mistakes in the analysis methods used. This makes human validation essential at every stage of the editorial process.

According to Gil, drawing attention to this principle is key to preserving the very essence of scientific research. She explained that the aim of the policy is to prevent the academic process from becoming "a self-contained cycle in which papers are both written and reviewed by machines", leading to a loss of critical judgement. As a result, authors must bear full responsibility for every stage of the research, writing and publishing process, while those in charge of reviewing their work may not delegate their critical judgement and must protect the confidentiality of the editorial process and the value of peer review.

Ensuring quality, rights and privacy

The policy also places a particular focus on the legal and ethical issues surrounding the use of AI. Authors must ensure they have the necessary rights to use their chosen AI tools and check the terms of use governing both the content they upload and the resulting output. They are therefore advised not to allow uploaded content to be used for training AI models and to use materials published in open access or for which they hold the appropriate rights wherever possible.

Under the policy, authors must also ensure that AI-generated content does not infringe third-party copyright or image rights and that the UOC will be able to use it in its publications.

In relation to data protection, the policy states that they must avoid entering into AI systems content that may include personal data or any information that may make an individual identifiable, and it recommends configuring the AI systems with the settings that provide the greatest guarantees of privacy.

A document designed to evolve

The UOC treats this policy as a dynamic document, to be updated in line with the constant evolution of both AI and its regulatory framework. In view of this, the university intends to review it on a regular basis and amend it as necessary to reflect changes in technology, the law and society.

This focus on adaptability is also one of the document's most notable traits. Martín explained that "it has been conceived as a key editorial policy to help the authors and reviewers who contribute to our journals to use AI in an ethical and responsible manner, and it will be regularly reviewed by the university to bring it into line with technological, legal and social changes."

The UOC's aim with this initiative is to provide a clear framework for both authors and reviewers of its journals, increasing transparency, quality and scientific integrity at a time when artificial intelligence has become an integral part of research practice. Rather than restricting the use of these tools, the policy establishes guidelines to ensure that they are used in a responsible and ethical manner and always under human supervision, upholding the principles of academic publishing.