The advancement of artificial intelligence (AI) has been accompanied by a global regulatory race. As AI systems scale, countries worldwide are stepping up efforts to establish rules and boundaries for the technology.

Within this global context, Latin American nations have also made strides in discussions on how to regulate AI. The region has introduced legislative proposals representing diverse approaches, reflecting both an ambition for technological development and concerns regarding social impacts.

A key focus of this debate lies at the intersection of AI and copyright, particularly concerning generative AI. On one hand, discussions center on the use of protected works to train AI models and the potential mechanisms to remunerate authors whose content is utilized. On the other hand, questions arise regarding the authorship of AI-generated creations produced with minimal or no human intervention.

We conducted an exploratory mapping of how Latin American countries are seeking to regulate AI. The analysis covers federal bills currently under debate that propose general regulatory frameworks, as well as enacted legislation where applicable. The goal is to determine whether – and in what manner – copyright considerations are integrated into these proposals.

1. Countries that include copyright and intellectual property in AI regulation

Brazil 

In Brazil, Bill No. 2,338/2023, which was passed by the Senate and is currently under review in the House of Representatives, dedicates a specific section to copyright. The current text requires AI developers to disclose which works were used to train their models. This information must be made available in an easily accessible public document.

The bill permits the use of protected works for text and data mining (TDM) provided that access is lawful, non-commercial, and strictly limited to what is necessary for scientific research. For-profit entities cannot benefit from this exception.

Rightsholders may opt out of having their works used in AI systems and can challenge such use even after the model has been trained. Furthermore, the bill establishes mandatory remuneration for rightsholders when their works are used, with compensation scaled according to company size and economic impact. It also provides for collective licensing over the use of works and protects image, voice, and identity, requiring AI systems to respect personal rights.

Colombia

In 2025, Bill No. 043/2025 was introduced in Colombia to regulate AI. The bill requires AI systems to respect copyright, related rights, image rights, and intellectual property, making express consent from the rightsholder or a valid legal basis mandatory to use protected works, phonograms, performances, broadcasts, or other protected content.

The text provides that the government will promote accessible and flexible licensing models to balance rights protection with value creation. However, the bill does not specify remuneration models, focusing on the obligation to protect copyright without defining how.

Bill No. 293/2024, which focuses specifically on the relationship between AI and copyright, is also progressing through the Senate. It proposes rules regarding the use of protected works in AI training, exceptions for text and data mining (TDM) in scientific research, and collective management frameworks for authorizing and remunerating the use of musical and audiovisual works.

Peru

In 2023, Peru enacted Law No. 31814, promoting the use of artificial intelligence for the country’s economic and social development. In 2025, the government approved its implementing regulations. The text establishes guiding principles for AI development, including respect for copyright and neighboring rights.

Under these regulations, the moral and economic rights of authors, as well as applicable international treaties, must be respected across all stages of AI system development, implementation, and use. Nevertheless, the regulations do not establish specific rules regarding licensing, consent, or remuneration for using protected works in AI training. Copyright protection is treated primarily as a guiding principle.

El Salvador

In 2025, El Salvador enacted Legislative Decree No. 234 (the Artificial Intelligence and Technologies Promotion Law), aimed at stimulating innovation and protecting AI developers.

The text assigns all intellectual property resulting from AI systems, algorithms, or models, including patents, copyright, trademarks, and trade secrets – exclusively to developers.

Regarding data use for training AI models, the law permits the use of publicly accessible data (content made publicly available by its owner) for research and development, provided that intellectual property rights and third-party privacy are respected.

Chile

Bill No. 16821-19, which seeks to regulate AI in Chile, was passed by the Chamber of Deputies in 2025 and is currently progressing through the Senate.

In an earlier version, the text contained articles addressing copyright, including provisions that required transparency regarding data used in AI training and recognized text and data mining (TDM) for research purposes as a permissible practice – exempting developers from obtaining authorization or providing remuneration to rightsholders.

However, in September 2025, the Chamber’s Committee of the Future voted to strike the articles dealing with these matters. The decision was driven by the difficulty of reconciling conflicting interests between AI developers, who advocated for open access to large volumes of data, and copyright holders, who were concerned with protecting their works. Disputes over the use of works and data in AI systems will continue to be resolved under the current Intellectual Property Law, leaving courts to decide cases on an individual basis (Pizzoleo, 2025; Ortiz, 2025).

Argentina

Argentina has multiple bills currently under consideration aimed at creating regulatory frameworks for AI. Bill 2130-D-2025 proposes a legal framework for AI focused on risks, rights, and innovation. It mentions intellectual property in the context of potential future disputes, but does not address copyright or the use of protected works in AI training.

Another notable proposal is the “Turing Law” (Bill 1013-D-2024), which adapts the Argentine legal system to AI. It distinguishes between human, AI-assisted, and synthetic (AI-generated) works. Under this proposal, synthetic works would receive no copyright protection, while AI-assisted works would be protected only to the extent of the human contribution. It does not propose remuneration for authors whose works are used in training datasets.

Other pending bills include S-0511/2025 (risk-based regulation), S-0556/2025 (creating the National AI Institute), and S-2573/2024 (ethical principles and oversight), none of which address copyright in training or compensation for the use of protected works.

Panama

Bill No. 181/2025 is pending its first debate in the National Assembly. Among its core principles, the text establishes the protection of intellectual property. It mandates that AI systems maintain traceability of sources consulted during information processing, thereby safeguarding copyright, patents, trademarks, trade secrets, and other rights.

The bill also creates obligations for users of AI systems. Users must adhere to the principles of the law and disclose when AI is used in works that incorporate data, images, sounds, trademarks, or other content protected by copyright. Disclosing the use of AI does not relieve the individual or entity of liability in the event of an intellectual property infringement. The text does not establish rules regarding the authorship of AI-generated content, the use of protected works to train AI models, or remuneration for the use of such works.

Mexico

Several bills currently under consideration propose regulatory guidelines. One such proposal introduces a Law for the Ethical Regulation of Artificial Intelligence and Robotics. The text outlines general principles, but does not directly address copyright or the use of protected works in AI training.

In contrast, the bill introduced by Senator Ricardo Monreal Ávila mandates that developers and providers of AI systems based on large language models (LLMs) may only use third-party content subject to prior agreements with intellectual property rightsholders.

In addition to legislative proposals, in 2025 the Supreme Court of Justice of the Nation established binding precedent by ruling that works created exclusively by AI cannot receive copyright protection in the country (Mexico, 2025).

Venezuela

Venezuela approved its Artificial Intelligence Bill in a first discussion in the National Assembly in 2024, with a second vote scheduled for 2025. The proposal focuses on promoting scientific and technological development, establishing an oversight body, and drafting a code of ethics to guide the responsible use of AI.

Although the text mentions intellectual property as a guiding principle, it lacks clear definitions regarding copyright or the ownership of AI-generated creations (Grupo Lazarus, 2025).

Dominican Republic

In the Dominican Republic, Initiative 00495/2025-SE is currently before a legislative committee. The text addresses general principles concerning fundamental rights, focusing on privacy, security, and non-discrimination, without specifically dealing with copyright.

The bill provides for the protection of the intellectual property of AI technologies developed within the country and the national recognition of local innovations. Intellectual property also appears as legal grounds to challenge decisions made by AI systems that infringe upon fundamental rights (Senado de la República Dominicana, 2025).

Uruguay

In 2024, Uruguay adopted a Recommendation Report for AI Regulation aimed at ethical development, human rights protection, and fostering technological innovation, within the framework of the National AI Strategy 2024–2030. The document recognizes that many AI-related activities require the large-scale use of copyrighted works – both for text and data mining (TDM) and for training generative models. Consequently, it recommends updating the Copyright Law (Law No. 9,739) to include specific exceptions permitting the use of works for computational analysis, provided that such use does not conflict with a normal exploitation of the work nor unreasonably prejudice rightsholders.

Furthermore, it suggests initiating a public debate regarding works generated exclusively by AI, without human intervention, and AI-assisted works. While there is no consolidated stance on who should be recognized as the author, the report recommends evaluating mechanisms for authorship attribution, mandatory labeling of AI-generated content, and potential economic compensation for rightsholders whose works serve as the foundation for these creations.

2. Countries with AI bills that do not mention copyright

Bolivia

Bolivia made progress in AI regulation with Bill No. 178/2024-2025, which was passed by the Senate in 2025 and sent to the Chamber of Deputies for review.

For this project – the Law for the Promotion, Management, and Use of Artificial Intelligence – only the explanatory note (exposición de motivos) has been made public. The full text containing the proposed articles is not currently available (Correo Del Sur, 2025).

The explanatory statement seeks to balance technological innovation with the protection of fundamental rights, acknowledging associated risks. The bill also proposes establishing a regulatory authority with sanctioning powers (Cámara de Senadores Bolivia, 2025). The statement does not address copyright or intellectual property. The absence of the complete statutory text precludes a definitive analysis of its provisions.

Ecuador

Ecuador has several AI bills under consideration with varying scopes (Guzmán, Guachilema, Cataña, 2025). The primary proposal is the 2024 Organic Bill on the Regulation and Promotion of Artificial Intelligence. It proposes a legal framework based on risk classification, but does not address copyright or remuneration for the use of works in AI.

Additionally, the AI Promotion and Development Bill and the Organic Bill on Digital Utilization and AI for Children and Adolescents are currently progressing through the legislature, neither of which addresses the use of protected works, licensing, or remuneration.

To date, Ecuador appears to favor a technical and institutional approach to AI regulation, leaving copyright outside its scope.

Paraguay

In 2025, Paraguay introduced a Bill to the Senate to regulate and promote the creation, development, innovation, and implementation of Artificial Intelligence systems. The text does not address copyright, focusing instead on risk categorization and the requirement for reliable data in AI systems.

Costa Rica

Costa Rica has multiple AI bills currently under consideration (such as Bills 23771 and 23919), which remain controversial and appear far from legislative enactment (Guillen, 2025). The proposals focus on topics such as ethical regulation, human oversight, and data protection, but do not specifically address intellectual property or copyright.

3. Countries without AI bills

As of the time of this mapping, no general bills regulating artificial intelligence could be identified in the following countries: Cuba, Guatemala, Haiti, Honduras, and Nicaragua.

4. What does this landscape reveal about regulatory pathways for AI and copyright in the region?

The mapping of AI bills across Latin America indicates a protective stance toward copyright in AI training. This outcome is hardly surprising. In a previous study on global copyright laws (Ribeiro; Garrote, 2025), Brazil and most Latin American nations already exhibited restrictive regimes, with limited flexibilities for research and sparse provisions for text and data mining (TDM).

Comparative studies on TDM exceptions confirm this pattern. Many countries in the region adopt restrictive copyright approaches with minimal exceptions (Fiil-Flynn et al., 2022). Copyright legislation in Latin America has grown increasingly restrictive over time – a trend at odds with that of developed nations – which could negatively impact innovation, technological development, and research across the region (Schirru et al., 2024).

Against this backdrop, the AI regulation proposals currently under debate in Latin America appear to follow three main pathways:

1. Principles-based approach: The majority of pending bills mention copyright and intellectual property merely as general principles. In these cases, the AI law defers questions of authorship, the lawfulness of training uses, and remuneration to general copyright statutes or to the judiciary.
2. Targeted flexibilization movements: A small handful of countries introduce elements pointing toward greater flexibility. This is reflected in legislative initiatives that are beginning to discuss broader exceptions for research and technological development.
3. Detailed regulation within the AI law itself: A smaller fraction of proposals, such as Brazil’s, attempts to directly address the use of protected works in AI training by establishing requirements for prior authorization, source traceability, or specific licensing obligations.

Countries such as Brazil, Colombia, Peru, El Salvador, and Mexico already incorporate copyright references into their AI bills to varying degrees. Chile, on the other hand, saw specific provisions on the subject removed during the legislative process.

Conclusion

Overall, the evidence shows that most proposed legislation in the region has not yet comprehensively addressed the challenges that AI poses to the copyright regime.

This mapping was exploratory and non-exhaustive, based on information available up to October 2025. It did not evaluate the specific legislative stages of the bills or the political prospects of their passage.

The regulatory landscape remains dynamic, with proposals still pending or in early stages. A notable lack of uniformity persists across Latin American countries – there is no single path for regulating AI, highlighting a debate that is very much still taking shape.

Consolidated Table