Revista Verba Iuris

Copyright, research and use of artificial logic in Colombia^*

Derecho de autor, investigación y uso de la lógica artificial en Colombia

Marco Antonio Ruiz Nieves^**

Jhonny Alexander Marín Quesada ^***

^*Article is a research result derived from the project entitled Artificial Intelligence and Jurisdictional Function in Colombia: Constitutional Limits, Procedural Guarantees and Digital Peacebuilding, identified with the code INV-DNI-DER-BOG-GR05-PROY2026-28, linked to the Peace Research Group of the National Unified Corporation of Higher Education (CUN).

^**Lawyer, with a master’s degree in human Rights and International Humanitarian Law for litigation before international courts and tribunals from the Universidad Santo Tomás, a research professor and professor (associate) of the National Unified Corporation of Higher Education (CUN). CvLac: https://scienti.minciencias.gov.co/cvlac/visualizador/generarCurriculoCv.do?cod_rh=0001680997 ORCID: https://orcid.org/0000-0001-5963-5180

^***Lawyer and psychologist, with interdisciplinary training oriented to the comprehensive analysis of legal and social relations. Master’s degree in business labor law from the University of La Sabana. ORCID: https://orcid.org/0000-0001-5963-5180 email: jhonnymaqu@unisabana.edu.co

Abstract

The article examines the relationship between copyright and the use of artificial logic systems in textual production in Colombia, focusing on the whether technological mediation affects the legal concept of authorship. The study is structured around three distinct areas of analysis: the copyright regime, academic textual production, and legal textual production in the context of litigation. The first section examines the administrative doctrine of the National Copyright Directorate (Dirección Nacional de Derecho de Autor) and its restrictive interpretation of authorship, emphasizing that authorship is recognized exclusively in natural persons. The second section analyzes academic production as the outcome of an intellectual process that extends beyond the mere drafting of a text to encompass the broader construction of knowledge. From this perspective, technological tools may be understood as forms of instrumental support within the creative and scholarly process. The third section addresses the drafting of legal documents in which the value of the text lies primarily in the consistency of its legal reasoning and in the accurate verification of statutory and jurisprudential sources. Based on these distinctions, the article argues that artificial logic does not alter the legal foundations of copyright or displace human authorship. Rather it functions as a tool within the intellectual process, whose use requires rigorous oversight by the legal professionals, particularly with regard to the accuracy, reliability, and verification of the information employed.

Keywords

copyright law, authorship, artificial intelligence, artificial logic, writing, academic ethics.

Resumen

El artículo examina la relación entre el derecho de autor y el uso de sistemas de lógica artificial en la producción textual en Colombia, centrándose en determinar si la mediación tecnológica afecta el concepto jurídico de autoría. El estudio se estructura en torno a tres áreas de análisis diferenciadas: el régimen de derecho de autor, la producción textual académica y la producción textual jurídica en el contexto del litigio. La primera sección examina la doctrina administrativa de la Dirección Nacional de Derecho de Autor y su interpretación restrictiva de la autoría, destacando que esta se atribuye exclusivamente a las personas naturales. La segunda sección analiza la producción académica como resultado de un proceso intelectual que trasciende la mera redacción de un texto para abarcar la construcción más amplia del conocimiento. Desde esta perspectiva, las herramientas tecnológicas pueden entenderse como formas de apoyo instrumental dentro del proceso creativo y académico. La tercera sección aborda la elaboración de documentos jurídicos, en los cuales el valor del texto radica principalmente en la consistencia argumentativa y en la verificación de las fuentes normativas y jurisprudenciales. A partir de estas distinciones, el artículo sostiene que la lógica artificial no altera los fundamentos jurídicos del derecho de autor ni desplaza la autoría humana. Por el contrario, funciona como una herramienta dentro del proceso intelectual, cuyo uso requiere una supervisión rigurosa por parte de los profesionales del derecho, particularmente en lo relacionado con la exactitud, confiabilidad y verificación de la información utilizada.

Palabras clave

derecho de autor, autoría, inteligencia artificial, lógica artificial, escritura, ética académica.

Abstract

Modern reflection on autonomy and liberal ideology reveals a tension concerning the conception of political freedom present in the idealism of Rafael Uribe Uribe. From this perspective, the relationship between Immanuel Kant’s practical philosophy and the liberal thought of the Colombian leader Rafael Uribe Uribe is examined, with particular attention to the articulation between political freedom and republicanism. From a hermeneutic-comparative approach, it is argued that the Kantian demand for autonomy and Uribe Uribe’s defense of state intervention aimed at social justice do not constitute incompatible positions, but rather differentiated moments within the same normative structure. Uribe Uribe's social republicanism is interpreted as a historical mediation that translates the universal principle of autonomy into concrete institutional conditions and, in doing so, reconfiguries the concept of citizenship. This line of inquiry makes it possible to situated his thought within the genealogy of Latin American social constitutionalism and to provide a conceptual framework for contemporary debates on social, cultural and economic rights, as well as on the social rule-of-law state.

The philosophical reflection does not seek to exhaust the notion of freedom throughout the entirety of the Kantian corpus—a polysemic category that traverses transcendental cosmology, practical reason, and the doctrine of right—but rather confines itself specifically to the dimension of the autonomy of the will. Likewise, the reference to Rafael Uribe Uribe's idealism is not approached through a general ontological redefinition of the term, but rather through the specific reconfiguration of the political that the liberal leader articulated within the Colombian socio-legal context of the late nineteenth and early twentieth centuries. It is at the intersection of this Kantian normative directive and this proposal for political praxis that the central research problem is located.

Keywords

transcendental autonomy, social republicanism, political freedom, situational hermeneutics, Colombian liberalism.

1. Introduction

Recent technological developments have transformed the ways in which knowledge is produced, organized, and disseminated in academic and legal contexts. Within this landscape, algorithmic systems capable of processing large volumes of information and generating textual responses have emerged and are commonly grouped under the broader concept of artificial intelligence. However, a closer examination of how these systems operate suggests that they do not function as autonomous agents of creation. Rather, they are algorithmic systems designed by human beings that process and reorganize pre-existing information in response to specific instructions. From this analytical perspective, it may be more appropriate to understand them as systems of artificial logic that operate as tools within the processes of intellectual work.

The increasing use of these technologies has prompted considerable debate in academic and legal settings regarding their impact on textual production and on traditional principles of copyright law. It has been raised whether automated text-generation systems may transform the legal concept of authorship or alter the foundations of intellectual property protection. Such debates, however, are often conducted without adequately distinguishing between the different forms of writing found in academic and legal contexts, which may lead to imprecise assessments of the actual scope of the issue.

The analysis developed in this article begins by identifying three distinct contexts in which texts are produced within the contemporary legal field. The first corresponds to the traditional domain of copyright law, particularly with respect to literary works and other creations susceptible to registration with the National Copyright Directorate (Dirección Nacional de Derecho de Autor). At this level, the debate centers on intellectual authorship and the originality requirements governing copyright protection. The second context concerns texts resulting from academic research processes, in which the primary value of the work lies not merely in its written form but in the methodological and intellectual process underlying the production of scholarly knowledge. The third context involves documents produced in the professional practice of law, such as complaints, written submissions, and appeals filed before the different jurisdictions in Colombia, whose purpose is related to the argumentative construction oriented to the application of the law within a judicial or administrative proceedings.

Distinguishing among these three context allows for a more rigorous examination of the controversies surrounding generative technologies in the legal field. In the context of literary creations, the central issue concerns authorship and the protection of expressive form; in academic work, with the methodological organization of the knowledge produced; and in litigious practice, with argumentative consistency and the verification of the rules and judicial decisions invoked.

From this analytical framework, the research problem guiding this study is to determine how the use of artificial logic systems in these different forms of textual production should be legally interpreted within the Colombian context, particularly in relation to the copyright regime.

Accordingly, the study addresses the following research question: How does the use of artificial logic systems in the production of academic and legal texts relate to the Colombian copyright regime without altering the legal concept of authorship? The objective of this article is to analyze the relationship between copyright and the use of artificial logic systems in academic and legal textual production in Colombia through an examination of the international, Andean Community, and national regulatory frameworks, recent constitutional jurisprudence and the administrative doctrine developed by the National Copyright Directorate.

The study therefore examines technological mediation through two distinct but related questions: the attribution of authorship under copyright law and the attribution of intellectual or professional responsibility for the content incorporated into academic and legal texts. This distinction prevents the use of a technological tool from being automatically equated with machine authorship, plagiarism or the displacement of human responsibility.

2. State of the art

Contemporary debates on generative artificial intelligence and copyright have gained increasing prominence in international legal scholarship, particularly in relation to questions of authorship, originality, human creativity, and responsibility arising from the use of automated text-generation systems. In this context, a significant part of the recent literature has focused on determining whether outputs generated by artificial intelligence systems may qualify for copyright protection and on defining the legal significance of human intervention in technologically mediated creative processes.

From this perspective, Picht and Thouvenin (2023) argue that the contemporary intellectual property system remains centered on the figure of the human creator, despite the rapid development of generative artificial intelligence tools. According to the authors, technological change does not automatically displace the anthropocentric foundations of copyright law, since legal protection continues to be linked to the intellectual activity performed by natural persons. This discussion has contributed to the consolidation of a doctrinal approach that treats artificial intelligence not as an autonomous subject of rights but as a technological tool whose operation remains subject to human direction and control.

Similarly, Geiger (2024) argues that contemporary discussions concerning artificial intelligence and copyright should be approached from a perspective consistent with human rights. The author maintains that copyright protection is directly related to the dignity, personality and creative freedom. Accordingly technological developments should not result in the dehumanization of the intellectual property system. From this perspective, recent literature has increasingly questioned models based on the complete automation of creativity and has emphasized the need to preserve meaningful human intervention as the structural element of copyright protection.

Van Woudenberg, Ranalli, and Bracker (2024) develop a conservative stance regarding authorship in generative systems such as ChatGPT, arguing that these tools cannot be considered authors in the strict legal sense. The authors explain that authorship implies awareness, intention and the ability to assume responsibility for the content produced, none of which is present in automated probabilistic language-processing systems. This approach has been adopted by different doctrinal sectors to argue that artificial intelligence lacks autonomous capacity for legal imputation within contemporary copyright systems.

The term artificial logic, as proposed in this article, is not intended to replace technical expressions such as natural language processing systems, large language models or generative models. These terms refer to specific technological architectures, computational procedures, or functions. Artificial logic, by contrast, is used here as a legal-analytical category intended to emphasize the absence of consciousness, intentionality, and autonomous legal agency in such systems. Its purpose is therefore critical rather than taxonomic: it seeks to question the anthropomorphic implications associated with the term intelligence while preserving the technical distinctions established in computer science and specialized literature.

With regard to originality and hybrid creative processes, Mazzi (2024) examines the legal scope of prompts and instructions given to generative systems. The author argues that such instructions often operate as abstract ideas or guidelines that do not necessarily satisfy the threshold required for copyright protection. However, he also notes that certain levels of subsequent human intervention could constitute protectable creative manifestations depending on the verifiable intellectual participation present in the final work. This discussion has acquired special relevance within contemporary doctrine, because it shifts the debate from the simple use of artificial intelligence to the concrete assessment of human contribution within technologically mediated creative processes.

Similarly, Mazzi and Fasciana (2024) argue that contemporary debates cannot be reduced to an absolute opposition between human creativity and artificial production. According to the authors, current processes of textual production reveal complex interactions between human intellectual activity and algorithmic processing. They therefore propose examining these processes through hybrid models of creation. This perspective is consistent with the approach developed by Thongmeensuk (2024), who observes that the expansion of generative systems has transformed discussions related to algorithmic training, data mining, and massive use of protected content in artificial intelligence models.

Within the same doctrinal context, Lucchi (2024) examines ChatGPT as a paradigmatic case for assessing contemporary copyright challenges posed by generative systems. The author argues that the legal issue is not limited to determining ownership of AI-generated outputs but also encompasses the legal boundaries of algorithmic training based on copyright-protected works. This line of research has broadened the debate by drawing attention to the tension between technological innovation and the large-scale use of protected content in automated learning systems.

In academic and legal contexts, another important sector of recent literature has focused on the risks derived from the use of generative tools within scenarios of scientific production and judicial litigation. In this regard, Abdali, Anarfi, Barberan, and He (2024) argue that contemporary AI-generated-content detection systems rely on probabilistic methods rather than mechanisms capable of identifying with absolute certainty whether a generative model has contributed to a particular text. The authors explain that highly technical and specialized forms of language may produce false positives in automated detection systems, a situation that acquires relevance within academic and legal contexts characterized by highly standardized linguistic structures.

Dahl, Magesh, Suzgun, and Ho (2024), in turn, conducted one of the most relevant studies on so-called algorithmic hallucinations within the legal field. The authors identified that generative models have a significant tendency to produce non-existent normative and jurisprudential references that nevertheless appear authentic, because such systems operate through probabilistic language modeling rather than autonomous mechanisms of legal verification. These findings have generated particular concern in both litigation and legal academic production, due to the risk of incorporating false information into judicial writings and scientific documents.

Consistent with this line with this line of research, Magesh et al. (2025) analyzed specialized AI-based legal research tools and concluded that even systems specifically designed for legal research remain vulnerable to significant risks of algorithmic hallucination. The authors argue that the incorporation of advanced information-retrieval mechanisms does not entirely eliminate the possibility of error and does not replace the need for rigorous human verification of the sources relied upon. This set of research has consolidated a doctrinal line aimed at maintaining that artificial intelligence should operate as an instrumental support tool rather than as a substitute for the professional and intellectual responsibility of the legal operator.

In Colombian, discussions concerning artificial intelligence and copyright have begun to develop across regulatory, administrative and institutional domains. CONPES Document 3975 of 2019 established the national policy on digital transformation and artificial intelligence and recognized the need to strengthen institutional capacities in response to technological development and digital automation (National Planning Department, 2019). Subsequently, CONPES Document 4144 of 2025 further addressed issues related to algorithmic governance, human supervision, and ethical use of artificial intelligence systems in Colombian (National Planning Department, 2025). These actions demonstrate that the Colombian debate on artificial intelligence increasingly incorporates concerns related to responsibility, institutional oversight, and human supervision in the use of generative technologies.

Along the same lines, the Superior Council of the Judiciary (Consejo Superior de la Judicatura) issued Agreement PCSJA24-12243 of 2024, establishing guidelines for the respectful, responsible, safe, and ethical use of artificial intelligence tools within the Judicial Branch (Superior Council of the Judiciary, 2024). The agreement develops principles related to traceability, transparency and human oversight of the use of generative technologies by judicial officials. Likewise, the Colombian Constitutional Court, through Judgment T-323 of 2024, established that artificial intelligence tools cannot replace rationality or human responsibility within the jurisdictional function (Constitutional Court of Colombia, 2024).

This institutional framework was further developed through Order AC739-2026 of the Civil, Agrarian and Rural Cassation Chamber of the Supreme Court of Justice of Colombia (2026), which expressly addressed the risks derived from the uncritical use of generative tools in judicial litigation. The corporation examined a case supported by non-existent rules and apocryphal jurisprudential references produced by artificial intelligence, developing a technical and legal analysis on the so-called algorithmic hallucinations and on the non-delegable duty of verification that the legal professional retains in the face of the sources used within judicial writings. This institutional action shows that the Colombian debate on artificial intelligence and copyright is no longer limited exclusively to the creative or academic contexts but is directly projected on professional responsibility and the contemporary exercise of litigation.

3. Research Problem

The problem addressed in this article arises from the legal and academic tensions associated with the use of artificial logic systems in textual production in Colombia, particularly in the relation to the concept of authorship protected under the copyright law. Although these tools are increasingly used in the preparation of academic, creative and legal texts, the Colombian legal framework continues to associate authorship with human intellectual creation.

Within this context, the central problem is to determine how the use of artificial logic systems should be legally interpreted across different forms of textual production, especially in contexts where creative processes, academic research, and the elaboration of legal writings converge, where the instrumental use of generative technologies cannot be automatically confused with displacement of authorship, plagiarism or substitution of intellectual and professional responsibility of the human subject.

4. Hypothesis

This research does not formulate its hypothesis in terms of quantifiable variables, but rather through analytical categories consistent with the qualitative approach and the doctrinal legal method adopted in the study. Accordingly, the hypothesis is constructed on the basis of the normative, doctrinal, administrative, and institutional interpretation of copyright in relation to the use of artificial logic systems in academic and legal textual production in Colombia.

From this perspective, the study hypothesizes that the use of artificial logic systems in the elaboration of academic and legal texts does not alter the legal notion of authorship established under Colombian law, since intellectual creation continues to be linked to human intervention as a presupposition of authorial imputation and legal responsibility.

This hypothesis is developed through a set of interrelated analytical categories. The first is human authorship, understood as a structural principle of the Colombian and Andean copyright regimes, particularly under Andean Decision 351 of 1993 and Law 23 of 1982, where intellectual creation is attributed to the natural person. The second category is related to artificial logic as an instrumental tool, which makes it possible to examine the legal scope of generative systems within technologically mediated textual production processes. A third category is differentiated textual production, through which the study distinguishes among literary creations, products derived from academic research, and legal writings elaborated in the exercise of litigation, with the purpose of analyzing how the problem of authorship operates in each scenario. Likewise, the category of intellectual and professional responsibility makes it possible to examine the duties of verification, human control and verification of sources in the face of the use of generative technologies, especially within the legal and academic field. These categories guide the analytical development of the article and allow structuring the results derived from the research.

5. Methodological strategy

This research adopts a qualitative approach based on documentary and normative analysis within the framework of doctrinal legal research. Its purpose is to examine how the Colombian legal system addresses authorship in relation to the use of generative artificial logic systems across different contexts of textual production. Within this framework, the primary research instrument consist of a documentary and normative analysis matrix designed to organize, classify, and interpret doctrinal, regulatory, administrative and institutional sources related to copyright and artificial intelligence in Colombia.

Using this instrument, the research analyzes the Berne Convention, Andean Decision 351 of 1993, Law 23 of 1982 and Decree 1066 of 2015. The matrix also incorporates an analysis of the resolutions issued by the National Copyright Directorate between 2023 and 2025, allowing the identification of the interpretative criteria used against the registration of content produced through generative tools and against the requirement of human intellectual creation demanded by the Colombian authorship regime.

The instrument further enables the examination of the analytical categories identified in the study: human authorship, artificial logic as an instrumental tool, differentiated textual production and intellectual and professional responsibility. These categories are analyzed across three distinct contexts: literary works eligible for copyright protection, outputs derived from academic research and legal documents prepared in the course of litigation.

Accordingly, the methodological strategy allows the study of legal and administrative sources to be articulated with the analysis of the tensions identified around the use of artificial intelligence in contemporary textual production.

6. Human authorship, copyright and limits of the so-called artificial logic in the international, Andean and Colombian legal system

The legal analysis of copyright and artificial intelligence must begin with a structural distinction. Copyright does not protect every textual or visual output merely because it has been produced. Rather, it protects intellectual creation that are legally attributable to a person recognized as an author. This requirement remains central to international, Andean and Colombian copyright regimes, despite the expansion of automated content-generation systems. The decisive question, therefore, is not whether a technological system can generate language, images or formally complex sequences. The relevant legal question is whether the resulting expression can be attributed to a subject capable of holding moral and economic rights.

At the international level, the Berne Convention does not establish copyright protection on the basis of the machine or the technical process involved, but rather on the work and the rights of its authors. WIPO recalls that the Berne Convention protects the works and rights of authors and defines literary and artistic works as encompassing “every production in the literary, scientific and artistic domain, whatever may be the mode or form of its expression” (World Intellectual Property Organization [WIPO], 1979, art. 2). This formulation is not merely declarative. Its significance lies in the fact that the international copyright system protects expression while linking such protection to the legal category of author. This prevents the focus of the analysis from shifting toward an alleged subjective autonomy of the technological device. The issue, therefore, is not whether technology possesses a high degree of sophistication, but rather that copyright, even in its classical international formulation, was not designed to recognize creative subjectivity in entities lacking legal personality. Instead, it was conceived to protect works in relation to their authors.

This underlying principle is reinforced under Andean law, whose relevance to the Colombian legal system cannot be regarded as merely peripheral. Decision 351 of 1993 of the Andean Community is not only forms part of the applicable legal framework but also provides an express definition that is particularly significant for the issue examined in this article. Article 3 defines an “author” as the “natural person who carries out intellectual creation” (Andean Community, 1993). The legal significance of this definition is substantial. Original authorship is not attributed of any entity capable of producing an output, but specifically to a natural person—that is, to a human being. The wording of the provision therefore precludes the direct attribution of authorship to an artificial intelligence system or, in the conceptual terms adopted in this article, to a system of artificial logic.

Furthermore, Article 7 of Decision 351 defines the scope of copyright protection by establishing that “only the way in which the author's ideas are described, explained, illustrated or incorporated into the works is protected”, while expressly excluding ideas and ideological or technical content as such from protection.

This dual normative function determines, on the one hand, that the author is a natural person and, on the other, that the protection falls solely on the expressive form. In this way, the Andean regime prevents two common misunderstandings: recognizing authorship of a machine and assuming that any thematic or conceptual similarity constitutes an authorial infringement.

The same reasoning is reflected in the Colombian legal framework. Article 61 of the Constitution provides that “the State shall protect intellectual property for the time and through the formalities established by law”, so that the constitutional protection of intellectual property refers to concrete legal development and does not enable arbitrary conceptual expansions. It is precisely in this legal development that the relationship between creation and person becomes clearer.

Article 1 of Law 23 of 1982 establishes that “authors of literary, scientific and artistic works shall enjoy protection for their works”, while Article 30 provides that the author shall have a “perpetual, inalienable and inalienable” right over his work to claim the authorship of the work, to oppose distortions and to keep it unpublished or to disclose it. The hermeneutical consequence is unequivocal: moral law presupposes a relationship between work and personality that can only be predicated of a human person.

A system of artificial logic cannot claim authorship, n the personal sense contemplated by moral rights; it cannot experience harm to its authorial reputation, consent to or object to the mutilation or alteration of a work on moral grounds, or exercise a personal and intimate connection with the resulting expression. In other words, if moral rights are perpetual, inalienable and inalienable because they protect a projection of the author's personality, then artificial logic cannot be their original owner.

Colombian legislation itself also makes an essential distinction that helps avoid conceptual misunderstandings. Economic rights are may be assigned, licensed, or otherwise commercially exploited; however, their transferability does not eliminate or alter the original attribution of authorship. Accordingly, the question of artificial intelligence cannot be resolved simply by assuming that economic exploitation entails machine authorship, or that a platform becomes the rights holder merely because it “generates” an output. The transferability of economic rights does not alter the underlying structure of the copyright system: authorship originates in the human creator.

Indeed, the general term of protection applicable to economic rights in Colombia reinforces this personalization of the system, since Article 21 of Law 23 of 1982 provides that they correspond to the author during his lifetime and, after his death, for eighty years to those who have legitimately acquired them. This rule also helps explain why at the expiration of the patrimonial term the work enters the public domain: not because the historical authorship disappears, but because the patrimonial exclusivity ceases.

From the institutional perspective, the role of the National Copyright Directorate must also be considered. According to Decree 1066 of 2015, it is responsible for "the design, direction, administration and execution of government policies on copyright", to maintain the national registry of literary and artistic works and to exercise inspection and surveillance in the areas attributed to it by law. This institutional framework is not merely administrative or ancillary to the argument developed here. Rather, it is directly relevant to the thesis of this article because the registration and management of copyright in Colombia presuppose the identification of a legally recognizable rights holder rather than an algorithmic system lacking legal personality. Even when technological tools participate in the production of a particular output, the institutional structure of the Colombian copyright law remains anchored in the natural person as the original point of attribution.

Within this framework, Judgment T-323 of 2024 acquires particular interpretative significance. Circular PCSJC24-37 noted that the Constitutional Court addressed "the use of generative artificial intelligence tools in tutela proceedings" and ordered the dissemination of the judgment and the implementation of related training within the judicial branch.

The most significant normative contribution of the judgment, however, lies not merely in its dissemination but in the set of principles established by the Court. These include transparency, responsibility, privacy, the non-substitution of human rationality, rigor and verification, risk prevention, equality and equity, human oversight, ethical regulation, adaptation to good practices and continuous monitoring. The court defines transparency, for example, as the obligation to clearly demonstrate the use, scope and location of the tool in the decision; responsibility as the duty to understand impacts and verify the information provided; and, especially important for this article, the non-substitution of human rationality as the "ethical and legal impossibility of substituting the action and responsibility of the individual of the human species" in the management of judicial actions and decisions.

This principle of non-substitution of human rationality has a direct argumentative power for copyright. If in a field as sensitive as constitutional jurisdiction the Court understands that the tool cannot replace human rationality because responsibility, decision and motivation continue to be non-delegable, with all the more reason it must be maintained that artificial logic cannot displace the human subject in a regime that, like the authorial one, is based precisely on personal authorship, in moral imputation and in the legal ownership derived from creation. The ruling does not regulate copyright, but it does establish a thesis of substantive legal theory: technology can assist, accelerate, organize, suggest or complement; it cannot replace the human core of responsibility or legally relevant rationality. This criterion, transferred to copyright, strengthens the conclusion that artificial logic is not the original owner of copyright. It can intervene instrumentally in a process; she cannot become an author in the normative sense required by Decision 351 and Law 23 of 1982.

Judgment T-323 of 2024 also underlines the need for verification, risk anticipation and human supervision in the face of hallucinations, prejudices and inaccuracies. Moreover, the statement points out that the use of these tools demands "strict scrutiny of the sources, scope, restrictions, possibilities, shortcomings and risks."

From the perspective developed in this article, this reasoning also makes it possible to challenge another overly broad association that frequently appears in contemporary academic debate: the automatic equation between the use of AI and plagiarism. The sentence does not speak of authorial plagiarism, but it does impose a logic of ethical and controlled use of the tool. This suggests that the real problem is not the mere use of the system, but the absence of transparency, verification and human control over the result. Translated into academia, textual matching or automated editorial support cannot be confused with plagiarism outright. Plagiarism requires misappropriation of someone else's work or of a protected expressive form, while coincidence can be the effect of shared technical language, disciplinary commonplaces, concurrent sources or even the statistical functioning of the system itself. The indiscriminate use of the term "plagiarism" to describe any textual coincidence, in addition to being legally incorrect, is stigmatizing and obscures the difference between authorial infringement, textual similarity and instrumental use of a technology.

In this respect, the concept of artificial logic offers greater analytical precision than the expression artificial intelligence, insofar as it avoids anthropomorphic assumptions. Whereas the latter may suggest cognitive capacities analogous to those of human beings, the former highlights that these are algorithmic procedures that generate plausible responses without creative will or legal personality. In this way, Judgment T-323 of 2024 supports this reading by requiring human supervision and preserving decision-making rationality in the hands of the people.

The historical example of Bram Stoker (1897) and Nosferatu may briefly illustrate the actual scope of copyright protection and why its legal meaning should not be trivialized. The Library of Congress of the United States recalls that the creators of Nosferatu did not obtain authorization to adapt the novel Dracula and that they tried to modify elements of the work to avoid accusations of infringement; even so, the conflict led to the claim of Stoker's widow. This episode illustrates that copyright grants rights holders control over certain adaptations and derivative exploitations of a protected work while the relevant economic rights remain in force. At the same time, the example also demonstrates the other side of the system: when the patrimonial term expires, the work enters the public domain. In the Colombian context, given that Bram Stoker died in 1912 and that the general term of economic protection extends for the life of the author and eighty more years, Dracula is no longer subject to economic exclusivity.

This point is relevant for two reasons. First, it demonstrates that not every reuse of another person’s constitutes copyright infringement. Once a work has entered the public domain, the heritage barrier disappears. Second, even when a user employs an artificial logic tool to reinterpret or rewrite a work in the public domain, the eventual new creation is not attributed to the tool, but to the human person who guides the creative process and assumes intellectual responsibility for the result.

Taken together, international law, Andean regulations, the Colombian legal system and Judgment T-323 of 2024 support the conclusion that artificial logic cannot constitute an author or the original holder of copyright. Decision 351 requires the author to be a natural person, Law 23 of 1982 recognizes moral rights arising from the personal relationship between the author and the work; and constitutional jurisprudence preserves rationality and responsibility within the human sphere. Within this framework, the mere use of artificial logic systems does not, by itself, constitute copyright infringement. The legal problem consists of establishing whether there is appropriation of a protected expressive form, concealment of authorship or ignorance of the rules of citation and attribution.

7. The National Copyright Directorate’s administrative approach to artificial intelligence and authorship: an emerging doctrine with internal tensions

One of the most significant contributions of this research consists of the comprehensive examination of the resolutions and institutional opinions issued by institutional opinions between 2023 and 2025 concerning applications for the registration of works produced with assistance of artificial intelligence tools. This review shows that the DNDA has developed an administrative doctrine that is relatively consistent in its underlying premise, although not entirely uniform in its application. Its central position has remained constant: original authorship and the initial copyright ownership may be attributed only fall to a natural person, not to an algorithmic system, since the legal framework defines the author as "the natural person who carries out intellectual creation" and reserves moral rights for the individual responsible for that human creative activity (National Copyright Directorate [DNDA], 2024a).

In the institutional opinion of June 24, 2024, the DNDA stated that, in creations generated through artificial intelligence, "the intervention of the human intellect is only reflected in the instructions given to the artificial intelligence program but not in the execution or elaboration of the specific expression." On this basis, the Directorate concluded that such expressions do not fall within the category of copyright-protected works and that the person who provides the instructions cannot, on that basis alone, be considered the author of the resulting expression, since ideas themselves are not eligible for copyright protection (DNDA, 2024a).

The same reasoning was reiterated in an institutional response dated January 24, 2024, in which the Registry Office stated that providing instructions to a generative system does not, in itself, satisfy the legal concept of a protected work or make the program an author. It further indicated that expressions generated by artificial intelligence are not, as such, eligible for registration in the National Copyright Registry (DNDA, 2024b).

Against this background, the 2023 resolutions and a substantial portion of 2024 decisions reflect a markedly restrictive approach. Resolution 095 of 2023 denied the registration of several artistic works; Resolution 137 of 2023 rejected the proposition that interaction with ChatGPT was sufficient to establish authorship; Resolution 147 of 2023 denies the registration of stories described as "written by ChatGPT"; Resolution 185 of 2023 conclude that Midjourney prompts were insufficient to establish originality; and Resolution 421 of 2023 reaffirmed that, in the absence of direct human creative effort, the threshold for copyright protection is not satisfied (DNDA, 2023a; DNDA, 2023b; DNDA, 2023c; DNDA, 2023d; DNDA, 2023e).

The same approach can be observed in Resolutions 143 and 144 of 2024, which reject applications for registrations of AI-generated images on the grounds that the contributions described by the applicants were merely corrective or instrumental and did not demonstrate a sufficient human creative contribution. In Resolution 143, for example, the DNDA reasoned that adjustments involving brightness, contrast, saturation, or color temperature do not transform an image into an original human creation, because such intervention remain instrumental and do not establish that the final expressive form originated in the human intellect (DNDA, 2024c).

However, when this body of decisions is compared with the resolutions issued in 2025, a first internal tension within the DNDA’s administrative doctrine becomes apparent. Whereas in 2023 and 2024 the Directorate generally adopted a rigid interpretation according to which human authorship was displaced when the specific expression resulted from algorithm generation, by 2025 it began, albeit cautiously, to distinguishing between merely generative uses of artificial intelligence and its instrumental or auxiliary use within a human creative process.

This distinction had already been suggested in the institutional response issued in January 2024, when the Registry Office stated that "the use of the program as a tool within the process of creating a work" should be distinguished from the artificial intelligence program acting "as the creator of an artistic expression" (DNDA, 2024b). Nevertheless, this methodological distinction began to produce more visible consequences in the decisions issued during 2025.

The continuing tension is evident in resolutions 042, 058 and 152 of 2025. In these decisions, the DNDA maintained its traditional position that the mere introduction of instructions or words does not satisfy the originality requirement and, therefore that AI-generated outputs are therefore not eligible for registration when the individual merely provides commands and the system independently produces the final expressive form. This is affirmed, for example, in Resolution 042 when analyzing an audiovisual work whose plot, dialogue, scenery and music had been generated with artificial intelligence; and the same is reiterated in Resolution 058, in which it insists that human intervention is exhausted in the introduction of instructions while the concrete expression results from the algorithmic execution of the system (DNDA, 2025a).

This restrictive position began to soften with Resolution 154 of June 5, 2025. Although the DNDA rejected the inscription due to insufficient evidence, it admitted that "the images are the result of a hybrid creative process, where the AI tool functioned as a support", and maintained that the Andean system does not prohibit the use of technological instruments in creative activity, as long as human participation is decisive in the final configuration of the work (DNDA, 2025h).

This clarification is significant because the DNDA no longer characterizes artificial intelligence exclusively as an external element that necessarily disrupts the relationship between the author and the work. Instead, begins to recognize AI as a potential instrument within a creative process led by people. In this way, it departs from the idea of a total incompatibility between such technology and authorship and accepts, at least at the conceptual level, the possible existence of hybrid productions susceptible to protection.

This conceptual opening creates tension with Directorate’s earlier reasoning. In 2023 and 2024, the prevailing argument was that users merely contributed ideas, guidelines or commands and that the resulting expression therefore could not be attributed to them as authors. Resolution 154 of 2025 shifts the analysis from the mere existence of AI to the degree of subsequent or concurrent human intervention.

It is not a carefree contradiction in the logical sense of simultaneously affirming and denying the same thing, but it is a relevant methodological inconsistency: the DNDA changes the axis of assessment without expressly acknowledging that it is reformulating its own criterion. Before, the problem seemed to be solved by algorithmic mediation alone; then, the question becomes dependent on the intensity, demonstration and relevance of the human contribution within a hybrid process. This shift is not minor, because it modifies the answer to the central question of the article: whether the so-called artificial logic always excludes human authorship or if it only excludes it when it absorbs the substantive configuration of the final form.

A second tension concerns the treatment of prompts. As a general rule, the DNDA has maintained that prompts constitute ideas, descriptions or guidelines and, therefore, are not subject to protection. This thesis is consistently supported by Article 7 of Andean Decision 351 and by the authoralist tradition that distinguishes between idea and form of expression. DNDA itself expressly stated this in the June 2024 concept, stating that ideas are not susceptible to protection and that whoever gives instructions to an AI program cannot be considered the author of the resulting expressions.

In certain 2025 files, the entity also alters its way of evaluating the evidence. Instead of just considering prompts as mere abstract notions, it began to examine the sequence between instructions, editing, choice, and post-production. In Resolution 154, for example, the examination is aimed at establishing whether there were subsequent interventions of digital retouching, reconstruction, integration of components and editorial elaboration capable of introducing an original human contribution into the final image.

This difference in approach creates a further interpretative difficulty. If a prompt is always an unprotectable idea and the AI materializes the expression, the reasoning should lead, in all cases, to the automatic inadmissibility of the registration. But the DNDA does not act like this in 2025. Instead of closing the debate at the moment of the prompt, it shifts the evaluation to the existence of a subsequent human creative chain, which implies implicitly acknowledging that algorithmic mediation does not by itself eliminate any possibility of work. Again, it is not a frontal contradiction, but it is a doctrinal oscillation: the entity goes from a strongly exclusive model to another that admits an exception that is not fully systematized, that of the hybrid work.

A third friction arises in the audiovisual field. In resolutions 262 and 263 of 2024, as well as in resolutions 042 and 058 of 2025, the DNDA develops the concept of audiovisual work and the condition of director, producer and other creative participants; and indicates that Article 95 of Law 23 of 1982 requires the identification of specific persons in functions such as direction, script, music or animation. On this basis, it concludes that, if the audiovisual material was produced using artificial intelligence, it cannot be imputed to the applicant as his own work (DNDA, 2025b).

However, in these same decisions, the DNDA admits the collective nature of the audiovisual work, in which various subjects intervene, and the producer assumes a coordinating role, while the authors retain moral rights over their contributions. This setup would have allowed for a more detailed assessment of human intervention in contexts where some content is algorithmically generated and some is selected, organized, and edited by people. Even so, the entity does not fully develop this alternative and maintains a more restrictive approach than the one later adopted in relation to hybrid visual works.

Further tension can be identified between the general administrative doctrine and the specific practice of registration. In a response issued in January 2024, the Registry Office indicated that it does not have the competence to pronounce "in a general and absolute manner" on the registration of creations made with artificial intelligence and subsequently modified by people, noting that the valuation must be carried out "on a case-by-case basis" (DNDA, 2024d).

This institutional prudence contrasts with the categorical language used in June 2024 opinion and in several subsequent decisions, in which the DNDA seems to assume as an almost general rule that outputs generated by artificial intelligence do not constitute protected works and that individuals who merely formulate instructions cannot be recognized as authors.

As a result, the entity alternates between a position open to individual examination and a clearly generalizing one. This ambivalence diminishes the predictability of his doctrine, since it is not always evident when he formulates a general criterion of exclusion and when he rejects a case for insufficient evidence of human participation.

Accordingly, the central inconsistency in the DNDA's action does not consist in affirming or denying that artificial intelligence can be authorized, but in the fluctuation of the criteria used to determine when there is legally relevant human participation. Although the entity consistently maintains that the authorship corresponds to human beings, in some cases it considers the use of artificial intelligence sufficient to reject the registration, while in others it examines interventions such as correction, selection, assembly or restoration, or even indicates that each application must be assessed in a particular way. This instability reveals that administrative doctrine has not yet built a coherent framework for legally understanding technological intermediation in creation.

From the perspective adopted in this study, this fluctuation allows the problem to be delimited more precisely. The DNDA consistently maintains that an artificial intelligence system cannot be regarded as the original author or owner because the Colombian and Andean legal system are structured around the natural person as creator and around moral rights linked to human individuality. However, the Directorate has not yet offer a completely stable criterion for determining when the use of artificial intelligence remains at the level of mere technical assistance and when it truly displaces human creation. This is where the notion of artificial logic comes in handy: not to simplistically deny every technologically mediated work, but to show that the real legal problem is the degree to which the algorithmic system replaces, conditions, or interrupts the chain from human thought to the final expressive form.

The contradiction of position, therefore, is not a minor flaw in the drafting of resolutions; it is the symptom that the classical theory of authorship is being pressured by hybrid forms of production that the DNDA still describes with traditional categories, but without having yet constructed an entirely coherent dogma to resolve them.

Table I. Evolution of the administrative doctrine of the National Copyright Directorate on artificial intelligence (2023–2025)

Source: own creation based on the analysis of the resolutions of the National Copyright Directorate (2023–2025).

8. Analysis and results

The normative, jurisprudential and documentary analysis conducted in this study made it possible to identify a fundamental element for understanding the scope of technological intervention in contemporary intellectual production: not all forms of textual production within the legal field are governed by the same protection regime or subject to the same conceptual requirements under copyright law. This finding emerges from comparing the position adopted by the National Copyright Directorate with the Colombian constitutional framework and the professional practice of litigation.

The examination makes it possible to distinguish three types of written production that are often mixed in discussions of artificial intelligence: literary creations protected by copyright, works derived from academic research and legal texts generated in the context of litigation. Although these forms of writing may share certain discursive characteristics, they responds to different purposes, methods of production, and legal frameworks.

Literary works constitute the traditional core of copyright protection and include narrative writings, essays, and other written works that reflect an intellectual creation attributable to a natural person. In this context, the uniqueness and personal manifestation of the creator determine the recognition of moral and economic rights, so the use of technological tools must be examined in terms of the level of human intervention in the final expression.

Scientific and academic research outputs operate according to a different logic, although they can also be protected by copyright when they are specified in texts. In these cases, the main value lies in the methodological procedure by which knowledge is produced (problem statement, construction of the conceptual framework, application of methods and critical evaluation of results), intellectual tasks of the researcher. In this scenario, technological resources can intervene as auxiliary instruments without displacing human authorship of the research process. At this point, it is important to specify that documents that have technical language often tend to be confused with the use of generative artificial intelligence, as is the case with legal language, which is extremely rigid in its procedural projection. In some cases AI use detectors, they are not accurate in determining the percentage of AI use accurately.

Content detection tools generated by generative artificial intelligence present variations in results. Thus, the same text may result in different percentages, depending on the system used, which shows its lack of technical consistency. In this regard, recent studies warn that, as human and AI-generated texts become similar, even the most advanced detectors have significant limitations in their ability to differentiate (Abdali et al., 2024).

The study also recognizes a third category: legal writings prepared in the exercise of litigation, such as claims, responses, briefs, opinions, allegations or challenges. Although they may contain complex reasoning, their main purpose is not literary creation or scientific research, but the defense of legal positions within judicial or administrative proceedings.

This distinction is fundamental for research because it demonstrates that technological participation in the legal field must be assessed with its own parameters. In the litigation, what is decisive is not the literary originality of the text, but the consistency of the legal reasoning, the appropriate regulatory interpretation and the verification of the cited jurisprudential sources. For this reason, legal documents can integrate technological tools to facilitate the organization or argumentative ordering, provided that the legal professional confirms the accuracy of the information used.

Based on these elements, the results of the analysis show that the use of artificial intelligence or artificial logic systems in the substantiation of legal texts cannot be evaluated under the same criteria applicable to literary works or scientific research products. While in the first two cases the debate focuses on the intellectual authorship and originality of the expression, in the field of litigation the central issue is related to the reliability of the legal information used and to the professional responsibility of the lawyer in the face of the veracity of the rules and precedents cited.

Accordingly, the findings indicate that the use of technological tools in the preparation of legal documents is compatible with professional legal practice, provided that the legal professional rigorously verifies the authenticity, accuracy, and relevance of the statutory, jurisprudential, and doctrinal sources used. This element directly responds to the problem question of the study, since it shows that the main challenge in the use of generative technologies within the legal field does not lie in the substitution of the intellectual activity of the jurist, but in the risk of uncritical use of unverified information.

9. Conclusion

The research demonstrates that the Colombian legal system preserves a conception of authorship linked to human intellectual creation, even in the context of the rapid expansion of generative artificial logic tools. The review of the Berne Convention, Andean Decision 351 of 1993, Law 23 of 1982 and the administrative doctrine developed by the National Copyright Directorate shows that copyright protection continues to be structured around the natural person as the subject of legal imputation. This finding directly addresses the objective of analyzing the notion of authorship within the Colombian and Andean copyright regimes in the context of generative technologies.

Similarly, the analysis of the resolutions issued by the National Copyright Directorate between 2023 and 2025 revealed an interpretative approach that, although initially restrictive regarding content produced with artificial intelligence, progressively began to recognize scenarios of human participation within hybrid creative processes. The documentary review carried out made it possible to notice that the Colombian administrative authority does not recognize artificial intelligence as the original owner of copyright; however, it admits the possibility of valuing human intellectual intervention when there is verifiable creative direction. This finding responds to the objective related to the examination of the administrative criteria used in Colombia to interpret the relationship between artificial intelligence and copyright.

The study also distinguished among the three contexts of textual production examined throughout the article: literary creations susceptible to authorial protection, products derived from academic research and legal writings prepared in the exercise of litigation. This distinction demonstrates that the impact of artificial logic does not operate uniformly within all forms of written production, because each one responds to different purposes and different levels of intellectual responsibility. In this way, the analysis addresses the objective of determining how generative technologies operate across different forms of textual production.

From another perspective, the analysis established that one of the principal risk associated with the use of generative systems within the legal field is not exclusively related to the loss of authorship, but to the phenomenon of the so-called algorithmic hallucinations and the use of unverified information within judicial and academic writings. In this regard, the examination of Judgment T-323 of 2024 and Order AC739-2026 of the Civil, Agrarian and Rural Cassation Chamber of the Supreme Court of Justice of Colombia made it possible to identify that the legal professional retains a non-delegable duty to verify the normative, doctrinal and jurisprudential sources used in the litigation. This finding directly addresses the objective of examining the intellectual and professional responsibility associated with the use of artificial intelligence in legal and academic contexts.

In light of these findings, the study concludes that artificial logic may operate as an instrumental support tool within academic and legal processes, provided that its use remains subject to meaningful human oversight, rigorous verification of the sources employed and intellectual direction by the person responsible for the resulting content. From this perspective, the analysis developed within the article shows that the contemporary challenge is not to prohibit the use of generative technologies, but to legally delimit their scope and preserve human responsibility as the structural core of copyright and the professional practice of law.

Referencias Bibliográficas

Abdali, S., Anarfi, R., Barberan, C. J., & He, J. (2024). Decoding the AI pen: Techniques and challenges in detecting AI-generated text. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (pp. 6428–6436). Association for Computing Machinery. https://doi.org/10.1145/3637528.3671463

Andean Community. (1993). Decision 351 of 1993: Common regime on copyright and related rights. https://www.comunidadandina.org/StaticFiles/DocOf/DEC351.pdf

Congress of the Republic of Colombia. (1982). Law 23 of 1982 on copyright. https://www.funcionpublica.gov.co/eva/gestornormativo/norma.php?i=3431

Constitutional Court of Colombia. (2024, August 2). Judgment T-323 of 2024 (Case T-9.301.656). https://www.corteconstitucional.gov.co/Relatoria/2024/T-323-24.htm

Dahl, M., Magesh, V., Suzgun, M., & Ho, D. E. (2024). Large legal fictions: Profiling legal hallucinations in large language models. Journal of Legal Analysis, 16(1), 64–93. https://doi.org/10.1093/jla/laae003

Geiger, C. (2024). Elaborating a human-rights-friendly copyright framework for generative AI. IIC—International Review of Intellectual Property and Competition Law, 55(7), 1129–1165. https://doi.org/10.1007/s40319-024-01481-5

Lucchi, N. (2024). ChatGPT: A case study on copyright challenges for generative artificial intelligence systems. European Journal of Risk Regulation, 15(3), 602–624. https://doi.org/10.1017/err.2023.59

Magesh, V., Surani, F., Dahl, M., Suzgun, M., Manning, C. D., & Ho, D. E. (2025). Hallucination-free? Assessing the reliability of leading AI legal research tools. Journal of Empirical Legal Studies, 22(2), 216–242. https://doi.org/10.1111/jels.12413

Mazzi, F. (2024). Authorship in artificial intelligence-generated works: Exploring originality in text prompts and artificial intelligence outputs through philosophical foundations of copyright and collage protection. The Journal of World Intellectual Property, 27, 410–427. https://doi.org/10.1111/jwip.12310

Mazzi, F., & Fasciana, S. (2024). Video kills the radio star: Copyright and the human versus artificial creativity war. The Journal of World Intellectual Property, 27(3), 341–365. https://doi.org/10.1111/jwip.12304

National Directorate of Copyright. (2023a, March 13). Resolución n.º 095 de 2023 [Resolution No. 095 of 2023]. Ministry of the Interior. https://www.derechodeautor.gov.co/es/resoluciones

National Directorate of Copyright. (2023b, May 16). Resolución n.º 137 de 2023 [Resolution No. 137 of 2023]. Ministry of the Interior. https://www.derechodeautor.gov.co/es/resoluciones

National Directorate of Copyright. (2023c, May 2). Resolución n.º 147 de 2023 [Resolution No. 147 of 2023]. Ministry of the Interior. https://www.derechodeautor.gov.co/es/resoluciones

National Directorate of Copyright. (2023d, June 14). Resolución n.º 185 de 2023 [Resolution No. 185 of 2023]. Ministry of the Interior. https://www.derechodeautor.gov.co/es/resoluciones

National Directorate of Copyright. (2023e, December 1). Resolución n.º 421 de 2023 [Resolution No. 421 of 2023]. Ministry of the Interior. https://www.derechodeautor.gov.co/es/resoluciones

National Directorate of Copyright. (2024a). Inteligencia artificial: Autoría y titularidad [Artificial intelligence: Authorship and ownership]. Ministry of the Interior. https://www.derechodeautor.gov.co/es/pronunciamientos-sobre-ia

National Directorate of Copyright. (2024b, April 30). Resolución n.º 143 de 2024 [Resolution No. 143 of 2024]. Ministry of the Interior. https://www.derechodeautor.gov.co/es/resoluciones

National Directorate of Copyright. (2024c, April 30). Resolución n.º 144 de 2024 [Resolution No. 144 of 2024]. Ministry of the Interior. https://www.derechodeautor.gov.co/es/resoluciones

National Directorate of Copyright. (2024d, January 24). Respuesta a petición sobre inteligencia artificial y derecho de autor, radicado 1-2023-124935 [Response to a request concerning artificial intelligence and copyright]. Ministry of the Interior. https://www.derechodeautor.gov.co/es/pronunciamientos-sobre-ia

National Directorate of Copyright. (2025a, January 23). Resolución n.º 042 de 2025 [Resolution No. 042 of 2025]. Ministry of the Interior. https://www.derechodeautor.gov.co/es/resoluciones

National Directorate of Copyright. (2025b, February 10). Resolución n.º 058 de 2025 [Resolution No. 058 of 2025]. Ministry of the Interior. https://www.derechodeautor.gov.co/es/resoluciones

National Directorate of Copyright. (2025c, February 17). Resolución n.º 069 de 2025 [Resolution No. 069 of 2025]. Ministry of the Interior. https://www.derechodeautor.gov.co/es/resoluciones

National Directorate of Copyright. (2025d, March 14). Resolución n.º 079 de 2025 [Resolution No. 079 of 2025]. Ministry of the Interior. https://www.derechodeautor.gov.co/es/resoluciones

National Directorate of Copyright. (2025e, April 30). Resolución n.º 133 de 2025 [Resolution No. 133 of 2025]. Ministry of the Interior. https://www.derechodeautor.gov.co/es/resoluciones

National Directorate of Copyright. (2025f, June 5). Resolución n.º 152 de 2025 [Resolution No. 152 of 2025]. Ministry of the Interior. https://www.derechodeautor.gov.co/es/resoluciones

National Directorate of Copyright. (2025g, June 5). Resolución n.º 153 de 2025 [Resolution No. 153 of 2025]. Ministry of the Interior. https://www.derechodeautor.gov.co/es/resoluciones

National Directorate of Copyright. (2025h, June 5). Resolución n.º 154 de 2025 [Resolution No. 154 of 2025]. Ministry of the Interior. https://www.derechodeautor.gov.co/es/resoluciones

National Planning Department. (2019). CONPES Document 3975: National policy for digital transformation and artificial intelligence. https://colaboracion.dnp.gov.co/CDT/Conpes/Econ%C3%B3micos/3975.pdf

National Planning Department. (2025, February 14). CONPES Document 4144: National artificial intelligence policy. https://colaboracion.dnp.gov.co/CDT/Conpes/Econ%C3%B3micos/4144.pdf

Picht, P. G., & Thouvenin, F. (2023). AI and IP: Theory to policy and back again—Policy and research recommendations at the intersection of artificial intelligence and intellectual property. IIC—International Review of Intellectual Property and Competition Law, 54(6), 916–940. https://doi.org/10.1007/s40319-023-01344-5

Political Constitution of Colombia. (1991). Political Constitution of Colombia. http://www.secretariasenado.gov.co/constitucion-politica

Presidency of the Republic of Colombia. (2015). Decree 1066 of 2015: Single regulatory decree of the administrative sector of the interior. https://www.funcionpublica.gov.co/eva/gestornormativo/norma.php?i=76835

Stoker, B. (1897). Dracula. Archibald Constable and Company.

Superior Council of the Judiciary. (2024). Agreement PCSJA24-12243 of 2024: Guidelines for the respectful, responsible, safe and ethical use of artificial intelligence in the Judicial Branch. https://actosadministrativos.ramajudicial.gov.co/GetFile.ashx?url=~%2FApp_Data%2FUpload%2FPCSJA24-12243.pdf

Supreme Court of Justice, Civil, Agrarian and Rural Cassation Chamber. (2026, February 13). Order AC739-2026 (Case No. 11001-02-03-000-2025-05324-00; Reporting Justice Martha Patricia Guzmán Álvarez). https://archivodigitalapi.cortesuprema.gov.co/share/2026/2/Sentencias/AC739-2026.pdf

Thongmeensuk, S. (2024). Rethinking copyright exceptions in the era of generative AI: Balancing innovation and intellectual property protection. The Journal of World Intellectual Property, 27, 278–295. https://doi.org/10.1111/jwip.12301

van Woudenberg, R., Ranalli, C., & Bracker, D. (2024). Authorship and ChatGPT: A conservative view. Philosophy & Technology, 37, Article 34. https://doi.org/10.1007/s13347-024-00715-1

World Intellectual Property Organization. (1979). Berne Convention for the Protection of Literary and Artistic Works: Paris Act of July 24, 1971, as amended on September 28, 1979. https://www.wipo.int/wipolex/en/text/283698

Este artículo se publica bajo los términos de la Licencia Creative Commons Atribución-No Comercial-SinDerivadas 4.0 Internacional (CC BY-NC-ND 4.0).