Introduction
Many researchers using text and data mining (TDM), an artificial intelligence (AI) technique that is key to advancing research, rely on access to freely available but often copyrighted works. This makes such researchers vulnerable to legal risks. While such materials may be widely accessible, concerns include preventing inadvertent copyright infringement and finding a way to reconcile this with the right to research.1 Rights to conduct and receive or access research have a strong fundamental rights justification, to the extent that its exercise is linked to the right to access information that is recognized in all normative human rights texts, therefore, the right to research can be seen as part of the right to freedom of expression, and it is more and more important as its exercise increasingly depends on digital tools and digital content. Ultimately, it is the capacity to access knowledge that allows researchers—and through them the whole society—to realize the rights to education, culture, and science, and therefore it is common that research activities are protected within the IP systems in forms of restrictions on the scope of exclusive rights or as limitations and exceptions.
Several emerging national AI strategies in Latin America focus on harnessing AI for innovation and economic development,2 with protective provisions for intellectual property (IP) being a key piece of this ecosystem. This conception of policy instruments that normalizes AI for innovation often ignores the need to balance the needs of innovation with a human rights perspective.3 A narrow focus on innovation can neglect the need to protect other rights (such as the right to research) that facilitate societies’ ability to enjoy the results of science. Protecting this right within an IP regime should be seen as a prerequisite for ensuring that AI delivers benefits for all.
Amid the rapid adoption of AI in the majority world,4 and specifically in Latin America, there is a complicated balance between IP rights, especially copyright regimes, and international human rights standards. The majority world refers to the fact that the worldwide population of developing countries in the Global South is larger than the population of postindustrial economies in the Global North. A vacuum that privileges IP rights at the expense of other rights has negative effects on the exercising of those other rights, such as the right we all have to access knowledge.
Exceptions to copyright law to allow the practice of TDM for research would be a starting point that would allow researchers to conduct their work without threats by IP protections; this proposal has a strong basis in academic and policy studies that highlight the users rights side within copyright regimes as even the World Intellectual Property Organization (WIPO) has admitted.5 The Latin American Civil Society Alliance for Fair Access to Knowledge (the Alliance) has proposed that countries in the region include TDM exceptions in their laws to better address the challenges that researchers in the region face.
As scientists increasingly use the techniques of computational analysis, legal access to these tools (or a lack thereof) can be a facilitator of (or an obstacle to) people’s ability to effectively exercise their right to science in the AI era. This is true worldwide, but there is a regulatory asymmetry between countries in the Global North and those in the Global South. The Global North has generally done a better job of adapting copyright rules to meet societal needs, including by granting exceptions for researchers where necessary to allow them, for example, to mine large amounts of text or disclose digital security vulnerabilities, with the understanding that such provisions will also benefit social and economic growth. This article explores how this issue has arisen in Latin America, and it highlights how debates over Latin American AI strategies can help address this regulatory asymmetry. A current proposal for a TDM exemption in Brazil and the Alliance’s version of such an exemption are the focus of the final section, in the hope that it will stimulate discussion in countries in the region and beyond.
AI and the Right to Research
The World Health Organization warned the public on January 6, 2020, about the coronavirus outbreak, whereas BlueDot, a Canadian digital health company, was able to detect early signs of the virus in Wuhan and sent an alert to its customers in December 2019. BlueDot was able to identify the outbreak thanks to its use of TDM techniques to mine and analyze large volumes of news reports and scientific publications. By relying on TDM to process large amounts of global content,6 most research projects related to COVID-19 have used openly accessible but often copyrighted works, such as news reports and scientific publications on the coronavirus family of viruses.7
TDM and AI are not synonymous, but TDM—as a technique of computational analysis—often plays a critical role in advancing research projects that rely on algorithms, machine learning, and the extraction of information. The use of TDM for research purposes is legal in countries such as Canada, Germany, Japan, the United States, and most of the countries in the Global North, which make up a minority of the world’s population. These countries have designed their legal frameworks to help them respond effectively to threats such as emerging pandemics. On the other hand, in Latin America, and in most other countries representing the majority world, the same use by researchers would face legal challenges.8 Owing to outdated and restrictive approaches to copyrights, these countries have not adapted their legal frameworks to have a balanced system where other rights—such as the right to research and thus the right to enjoy the results of human science—are available to all.
The countries with the most restrictive panorama for TDM are all developing countries, including Latin American countries in particular. In Latin America, no country provides specific exceptions to copyright for research activities based on data science (such as TDM). Only Ecuador mentions TDM activities in its legislation, but does not provide an exception to copyright, instead it includes an exemption to liability for library and archive officials.
The situation will have to change if these countries are to support their own competitive capacity in the new landscape following a diagnosis such as that in Professor Raquel Xalabarder’s 2023 report for WIPO: “What remains unquestioned is that TDM is a fundamental tool for the advancement of research activities and will redefine the way research is done, worldwide. Researchers may rely on machines to ‘read’ data (works) and find patterns, trends and correlations that may be valuable for the development of future research. Researchers who do not, may be at a competitive disadvantage.”9
Balancing Copyright Protections With a Human Rights Perspective
TDM techniques require making copies of large volumes of books, articles, images, databases, and other copyrighted materials and then subjecting them to text mining processes. Using such materials, even if they are freely available in a library or on the internet, may constitute copyright infringement and expose researchers to potential risks unless they can claim that such activity is permitted by law. This legal risk is fundamentally tied to where in the world the researcher lives.
Consider this scenario for a moment. “A team of biologists wants to train a machine-learning model using thousands of images of plants and animals from the internet, with the goal of creating an application that recognizes all types of plant and animal species. Is it legal?” This is one of the scenarios formulated and analyzed according to the laws in nineteen different jurisdictions in Latin America in a copyright risk assessment project entitled “¿Es Legal?” (Is It Legal?).10 The answer to this scenario is “no” in eighteen of the jurisdictions in the study, with the possible exception of Cuba, where the answer is uncertain. The answer to the same scenario analyzed under U.S. law is that this practice could be covered under the fair use doctrine, while in Europe, this practice would be legal under the TDM exceptions provided for in Articles 3 and 4 of the Copyright and Related Rights Directive for a Single Market.11
These differing responses illustrate how an unbalanced copyright system can be perceived as a legal risk and, in practice, can fail to consider international human rights standards in the legal implementation of any exceptions around the world, including TDM.12 While researchers in the minority world have the legal security to engage in these activities, this is not the case for researchers from Latin America and other parts of the majority world.
The Need to Face the Problem of Unbalanced IP Regulations
The disregard for international human rights standards when developing countries strike a balance in terms of IP systems is not new.13 Over the decades, the international IP system has developed minimum legal standards to address the rights of copyright holders, while leaving the balancing tools for protecting other rights—such as exceptions and limitations for the benefit of educators, for the sake of freedom of expression, or for other reasons—to national laws and interested parties.
This situation puts researchers in the majority world at a disadvantage because their governments have not established the conditions necessary, including to support research-enabling access to TDM techniques. This problem has been well documented and is now at a point where it is necessary for policymakers to address the inequality. A good place to start is to ensure that the national AI policies of countries in the Global South address the situation. If the majority world is to be an active producer of science that enables developing countries to benefit substantially from AI-generated scientific knowledge, their governments must protect the right to research and help make AI-enabled research easier, including through the use of legal exceptions and limitations that can make copyright systems more in balance with other human rights obligations.
Bridge the Gap by Bringing Human Rights Into the Debate
After the COVID-19 pandemic began, international calls for the world to address regulatory asymmetries in patent and copyright protections began to gain traction.14 Prospective reformers argued that the ability of researchers to use computational analysis techniques in their research pursuits, in particular TDM or machine learning, should not be determined by where in the world they are located.
In the case of copyright law, these calls to action bore fruit in 2023, when the leading international body responsible for international IP lawmaking, WIPO, adopted a new agenda on limitations and exceptions proposed by African countries on the Standing Committee on Copyright and Related Rights (SCCR). At WIPO, negotiations are conducted by regional groups of countries. Recently, the African Group has been promoting the human rights perspective and seeking to have this organization set standards for exceptions to copyright protections. The adoption of the African Group’s work program on limitations and exceptions means that WIPO will work to review and possibly create international instruments on exceptions regarding three priorities,15 the one on exceptions to facilitate research in the digital era will have TDM exceptions for researchers as a key discussion topic.
The interest in working on research exceptions related to TDM was also brought to another WIPO committee, the Committee on Development and Intellectual Property. This committee approved an investment of 500,000 Swiss francs (about $560,000) in a pilot project on TDM to support research and innovation in universities and other research-oriented institutions in Africa.16 The results of this exercise will support the agenda of the SCCR.
As the United Nations’ international lawmaking body for IP, WIPO’s recent moves to better integrate human rights concerns are welcome; this shift should have a positive impact on local debates about member states’ regulatory approaches.
AI Strategies in Latin America as a Way to Prioritize the Right to Research
In the international lawmaking landscape, there is a need to balance IP and human rights regimes. With respect to copyright law in particular, the situation is improving in WIPO as there is already a working program to address exceptions and limitations, including the right to science through the right to research and specifically a TDM exception. However, the results at WIPO take several years to advance, so relevant countries should also make related changes to their own national policies in parallel. Governments from global majority countries can prioritize deliberation and rulemaking within their national AI strategies including the topic in the policy roadmap for AI that (with some luck) may include the allocation of resources to support implementation.
Three years after the pandemic began, the Alliance conducted a review of the national AI strategies in nineteen Latin American countries.17 This analysis showed that six countries (Argentina, Brazil, Chile, Colombia, Peru, and Uruguay) had already developed a strategy, Mexico was in the process of developing one, and Ecuador was in the process of making a preliminary diagnosis. In these strategies, Argentina, Brazil, Chile, Colombia, and Mexico consider IP to be a relevant aspect for their respective plans, but only Brazil has included TDM as a specific issue to be addressed.
National AI strategies in Latin America are an opportunity to begin mitigating the divide between the majority world and the minority world in terms of TDM policies. Other countries in the region should emulate Brazil’s leadership example in addressing the need for an exemption for TDM as an AI-enabled technique for researchers in its national AI strategy.
Considering that no country in Latin America has a copyright exception for computational analysis techniques like TDM, these countries’ post-pandemic national strategies should reflect on the lessons learned and evaluate how the right to research can be ensured in this evolving era. If IP proves to be a critical factor in AI development, then establishing a well-balanced IP system would be imperative for harnessing AI as a catalyst for innovation and progress in the region.
Notably, only Brazil has undertaken this endeavor so far regarding its copyright regime. It is worthwhile to consider Brazil’s efforts to date in a bit more detail. The Brazilian Artificial Intelligence Strategy was approved in 2021 by the Ordinance of the Ministry of Science, Technology, and Innovations.18 The Brazilian proposal has been the only strategy of its kind in Latin America that included a brief mention to a TDM exception,19 and it was the first official document to raise the topic as a Brazilian regulatory priority, opening the regulatory window that will follow.
In March 2022, the Senate set up a Jurists’ Committee on AI to analyze three legal projects that were being discussed (5.051/2019, 21/2020, and 872/2021) with the goal of establishing foundations, principles, and guidelines for the development and application of AI in Brazil. The committee held public hearings later that month and opened public consultation to gather comments later in the year. Some of the comments included the need for a balanced IP system, including analysis by local academics on how to address TDM for research purposes.20 The committee published its final report in December 2022 with a new proposal for a bill that in its Article 42 includes an exception for TDM.21
After the requisite debates and discussions, the proposed Article 42 treats TDM as a research tool with an open approach. The proposed exception does not limit its scope by defining who is a researcher, but instead considers the institutions in which TDM can take place; however, it does set some limits on the activity. Article 42 is not only a copyright exception, as it also addresses data protection concerns (Section 3) and applies to other computational analysis activities with a narrower scope (Section 2).
Brazil’s Federal Senate is currently debating the draft bill (Bill 2338/2023).22 If it succeeds, Brazil will be the first country in the region to have debated and defined a regulatory approach to the right to research including TDM. The international and local processes in this case run parallel in time: while Brazil decided to incorporate these discussions in its national AI strategy and start the regulatory process, WIPO was defining the work plan for the right to research. This interplay between international and local policy development will provide interesting opportunities for cross-pollination in the coming years and will place Brazil in a good position to impact the international process.
It is important that WIPO is working to understand how best to balance the copyright system with human rights concerns, and especially to ensure the right to research for all. However, it is also important to give greater attention and guidance on this issue to local processes. To this purpose, the Alliance, while following the WIPO discussions, has been developing a proposal for a copyright exception that draws on the lessons described by experiences and research developed within the larger international A2K coalition.23 The Model Clause is intended to stimulate discussion in countries in the region that are engaged in developing or implementing a national AI strategy. Brazil’s experience will likely serve as an example to other nations; the Alliance’s proposal, on the other hand, is meant to serve as a model that can be revised and adapted to local realities.
In the report analyzing how the AI national strategies have evolved in the region, the Alliance included a proposal that has the basic components to provide a much-needed balance between copyright law and the right to research: it features an open approach to research purposes, not limited to TDM, and includes computational analysis more generally, is open to any investigative use, is open to all works, is compatible with fair uses, and includes a norm that protects against contractual and technical restrictions.24
Latin American governments can use this model clause to initiate much-needed national discussions on how to authorize computational analysis techniques for research purposes, particularly TDM and machine learning, which could help make AI a tool for innovation and economic development in these countries.
Conclusion
Digital security is a key concept in a digitally transforming society; it requires effective collaboration between the public and private sectors and even involves individuals, especially digital security researchers, who are required to disclose the vulnerabilities they identify in digital systems to keep them from being exploited in harmful ways. Such disclosures may include portions of copyrighted code or could include anti-circumvention actions forbidden under copyright laws.
Digital security researchers have faced the risk of copyright litigation for decades, but the call for countries to update their copyright systems to protect them came only ten years after the wave of national policies in this area. It was in a 2022 policy report on digital security that the OECD recommended that all countries create “safe harbors” for researchers—to protect their ability to investigate vulnerabilities—and recommended that the necessary adjustments be made to legal frameworks, including copyright law.25 Previously, in 2021, the OECD had noted that some of the activities performed by digital security researchers could be considered copyright infringement, creating a chilling effect, and noted that copyright legal risk particularly affects the jurisdictions of developing countries “For example, free trade agreements led to the adoption by some developing countries of inflexible anti-circumvention measures inspired by the first version of the US DMCA. These countries then rarely updated their frameworks to reflect subsequent improvements in the US law.”26
The lack of action to protect digital security researchers and facilitate their collaboration to prevent the exploitation of vulnerabilities to digital systems means that the legal risk for digital security researchers continues; this situation should not be repeated for AI. Considering the lesson learned in the field of digital security policies regarding persistent legal risks for researchers, including copyright provisions, and how the COVID-19 pandemic underscored the need to balance IP protections with international human rights standards, states should not miss opportunities to incorporate the right to science and particularly to address the much-needed exceptions for TDM as a research technique key in the development of AI.
The gap between developed and developing countries on copyright exceptions and limitations is well documented and is already part of the debate at WIPO. Such a divide deserves to be addressed by relevant countries at the domestic level. National AI strategies can serve to broaden national discussions on IP beyond their essentially economic approach to include the human rights implications of IP in terms of exercising other rights, such as the right to research. Countries that start domestic discussions on how to balance IP and human rights can later inform relevant international discussions on the topic and vice versa. This will help countries in need of reform to develop better regulatory capacity in this area.
Balancing IP protections with human rights concerns is a pressing issue that needs to be addressed in outdated national copyright laws. An increasing number of member states with better and more experience on this topic will help to improve international lawmaking on the topic. Otherwise, the outdated nature of Latin American countries’ legal frameworks will continue to expose Latin American researchers who use copyrighted materials with techniques such as TDM to unfair hurdles compared to their counterparts in the minority world.
Notes
1 Christophe Geiger and Bernd Justin Jütte, “Conceptualizing a ‘Right to Research’ and Its Implications for Copyright Law: An International and European Perspective,” American University Washington College of Law, July 2022, https://digitalcommons.wcl.american.edu/cgi/viewcontent.cgi?article=1079&context=research.
2 AI national strategies are usually supported by analysis such as the one presented in a 2018 report by the Inter-American Development Bank (IDB) that states that “AI is more than just a new technological trend. It is a unique hybrid of capital and labor that creates a completely new productive workforce that can learn by itself. In the medium term, the growth rate for the region’s GDP could go from 3% to 4% based solely on this factor. Almost 50% of that increase could be generated by an increase in productivity, as AI enables human workers to focus their efforts on the work that adds the most value. However, it is estimated that the impact of AI on the GDP of Latin America and the Caribbean (LAC) will be up to three or four times greater than in developed economies. This will make closing the development gap even harder than is already the case and it reveals the risks that come with delaying the process of creating the conditions necessary for consolidating AI and using it to increase productivity.” See Institute for the Integration of Latin America and the Caribbean, Planet Algorithm: Artificial Intelligence for a Predictive and Inclusive Form of Integration in Latin America (Washington, DC: Inter-American Development Bank, July 2018, p. 16, https://publications.iadb.org/es/publications/english/viewer/Integration_and_Trade_Journal_Volume_22_No._44_July_2018_Planet_Algorithm_Artificial_Intelligence_for_a_Predictive_and_Inclusive_form_of_Integration_in_Latin_America.pdf.
3 The historical isolation between human rights and IP in the international treaties can be explained by the fact “These treaties do refer to the protections granted to authors and inventors as “rights”. But the principal justification for these agreements lies not in deontological claims about inalienable liberties, but rather in economic and instrumental benefits that flow from protecting intellectual property products across national borders.” See Laurence R. Helfer, “Human Rights and Intellectual Property: Conflict or Coexistence?,” Minnesota Intellectual Property Review 5, no. 1 (2003): 47–61, https://scholarship.law.umn.edu/cgi/viewcontent.cgi?article=1399&context=mjlst. Likewise, a national strategy sees AI as merely a driver of economic growth. For its enthusiasts, AI is essentially a development promise. It is a key element in defining the new economic world order. Given similar drivers, it is not surprising that human rights are not prioritized when IP meets national AI strategies.
4 The “majority world” expression it is used to “highlight how the majority of the human population of the world lives in geographic regions—variously addressed as the “‘Developing World’”, the “‘Third World’,” or the “‘Global South”’—that are rendered and remaindered as passive peripheries of ostensibly global problems and developments. Rather than assuming that knowledge and innovations move out of the so-called centers of Europe and the United States to the rest of the world, thinking “‘from the majority world”’ means tracing out emerging forms of knowledge, innovation, and labor in former and still-colonized spaces and how it is often expropriated, extracted, and made invisible” as described by Sareeta Amrute, Ranjit Singh, and Rigoberto Lara Guzmán, “A Primer on AI in/from the Majority World: An Empirical Site and a Standpoint,” Data and Society Research Institute, September 14, 2022, p. 6, https://datasociety.net/library/a-primer-on-ai-in-from-the-majority-world.
5 Raquel Xalabarder, “Scoping Study on the Practices and Challenges of Research Institutions and Research Purposes in Relation to Copyright,” World Intellectual Property Organization (WIPO) Standing Committee on Copyright and Related Rights, October 17, 2023, https://www.wipo.int/edocs/mdocs/copyright/en/sccr_44/sccr_44_4.pdf.
6 Mark Prosser, “How AI Helped Predict the Coronavirus Outbreak Before It Happened,” Singularity Hub, February 5, 2020, https://singularityhub.com/2020/02/05/how-ai-helped-predict-the-coronavirus-outbreak-before-it-happened; and Corey Stieg, “How This Canadian Start-Up Spotted Coronavirus Before Everyone Else Knew About It,” CNBC, March 3, 2020, https://www.cnbc.com/2020/03/03/bluedot-used-artificial-intelligence-to-predict-coronavirus-spread.html.
7 Will Knight, “Researchers Will Deploy AI to Better Understand Coronavirus,” Wired, March 17, 2020, https://www.wired.com/story/researchers-deploy-ai-better-understand-coronavirus.
8 Sean Flynn, et al., Implementing User Rights for Research in the Field of Artificial Intelligence: A Call for International Action, European IP Review 2020, Issue 7 (short policy proposal on role of international limitations and exceptions).
9 Raquel Xalabarder, “Scoping Study on the Practices and Challenges of Research Institutions and Research Purposes in Relation to Copyright,” World Intellectual Property Organization (WIPO) Standing Committee on Copyright and Related Rights, October 17, 2023, p. 23, https://www.wipo.int/edocs/mdocs/copyright/en/sccr_44/sccr_44_4.pdf.
10 “Es Legal?: Nueva Base de Datos con Análisis de Casos Hipotéticos Sobre Excepciones y Limitaciones al Derecho de Autor en América Latina,” [Is It Legal? New Database With Analysis of Hypothetical Cases on Exceptions and Limitations to Copyright in Latin America], Acceso Justo al Conocimiento, February 28, 2023, https://accesoalconocimiento.lat/2023/02/28/es-legal-nueva-base-de-datos-con-analisis-de-casos-hipoteticos-sobre-excepciones-y-limitaciones-al-derecho-de-autor-en-america-latina.
11 The Directive (EU) 2019/790 of the European Parliament and of the Council of 17 April 2019 “on copyright and related rights in the Digital Single Market” and amending Directives 96/9/EC and 2001/29/EC can be found at https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CELEX:32019L0790
12 To be precise, a 2020 research report from American University showing a comparison of the general landscape of exceptions in copyright systems around the world explain what they called “the development gap” indicating that “We see little difference in the trends between civil and common law countries. The big difference we see is in countries of different development levels. Contrary to the political rhetoric of some that portray developing countries as strategic free riders, openness in copyright user rights is more the strategy of the north than the south. The high-income countries in our study have more open user rights in their laws, and the gap between them and developing countries has been growing since the early 1990s. See Sean Flynn and Michael Palmedo, “The User Rights Database: Measuring the Impact of Copyright Balance,” Social Science Research Network Scholarly Paper ID 3082371, 2019, p. 15, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3082371.
13 Originally, international legal instruments on human rights contained links to IP. However, over time, the two regimes, human rights and IP, evolved into separate legal domains and only recently they are reconnecting. See Helfer, “Human Rights and Intellectual Property.”
14 Flynn, et al., “Implementing User Rights for Research in the Field of Artificial Intelligence.”
15 The three priorities of the work plan proposed by the African Group is about the capacity of heritage institutions to preserve the content they host, the promotion of exceptions for the digital environment for online education and research and the possibility of broadening the scope of the Marrakesh Treaty, so that it supports uses not only for visually impaired people.
16 WIPO Secretariat, “Revised Pilot Project on Text and Data Mining (TDM) to Support Research and Innovation in Universities and Other Research-Oriented Institutions in Africa – Proposal Submitted by the African Group,” WIPO Secretariat, April 28, 2023, https://www.wipo.int/edocs/mdocs/mdocs/en/cdip_30/cdip_30_9_rev.pdf.
17 Charquero and Rangel, “Politcas de Inteligencia Artificial y Derechos de Autor en America Latina,” [Artificial Intelligence and Copyright Policies in Latin America]. It is important to note that even if the title of the report is translated as “copyright,” the Latin American legal tradition is derived from the French droit d’auteur (author’s rights). While the difference is not addressed in this paper, it is important to acknowledge that it exists.
18 For the full text in the original Portuguese, see Brazilian Ministry of Science, Technology, and Innovations, “Estrategia Brasileira de Inteligencia Artificial,” [Brazilian Artificial Intelligence Strategy EBIA], Brazilian Ministry of Science, Technology, and Innovations, 2021, https://www.gov.br/mcti/pt-br/acompanhe-o-mcti/transformacaodigital/arquivosinteligenciaartificial/ebia-diagramacao_4-979_2021.pdf. For a summary in English, see Brazilian Ministry of Science, Technology, and Innovations, “Summary of the Brazilian Artificial Intelligence Strategy EBIA,” Brazilian Ministry of Science, Technology, and Innovations, 2021, https://www.gov.br/mcti/pt-br/acompanhe-o-mcti/transformacaodigital/arquivosinteligenciaartificial/ebia-summary_brazilian_4-979_2021.pdf.
19 “One of the highlights in this topic concerns the need to include a new type of copyright limitation for text and data mining”, from the original: “Um dos pontos de destaque nesse tópico diz respeito à necessidade de se incluir um novo tipo de limitação aos direitos autorais, para mineração de textos e de dados” that can be found at “Estrategia Brasileira de Inteligencia Artificial,” [Brazilian Artificial Intelligence Strategy EBIA], Brazilian Ministry of Science, Technology, and Innovations, 2021, p. 18, https://www.gov.br/mcti/pt-br/acompanhe-o-mcti/transformacaodigital/arquivosinteligenciaartificial/ebia-diagramacao_4-979_2021.pdf.
20 Allan Rocha de Souza, “Direitos Autorais e Mineração de Dados e Textos no Combate à Covid-19 no Brasil,” [Copyright and Data and Text Mining in the Fight Against Covid-19 in Brazil], in Perspectivas e Desafios Informacionais em Tempos da Pandemia da Covid-19 [Information Perspectives and Challenges in Times of the Covid-19 Pandemic] 16, no. 2 (2020), https://revista.ibict.br/liinc/article/view/5536; and “Como Anda o Debate Sobre Inteligência Artificial No Brasil?,” [How the Debate About Artificial Intelligence Is Going in Brazil], IDB Autoral, June 20, 2023, https://ibdautoral.org.br/2023/06/20/3-como-anda-o-debate-sobre-inteligencia-artificial-no-brasil.
21 Senator Rodrigo Pacheco, “Projeto de Lei no. 2338, de 2023,” [Bill No. 2338, of 2023], Senate of Brazil, 2023, https://www25.senado.leg.br/web/atividade/materias/-/materia/157233.
22 Senator Rodrigo Pacheco, “Projeto de Lei no. 2338, de 2023,” [Bill No. 2338, of 2023], Senate of Brazil, 2023, https://www25.senado.leg.br/web/atividade/materias/-/materia/157233.
23 “About/Mission,” Access to Knowledge Coalition, https://www.a2k-coalition.org/about.
24 Patricia Díaz Charquero and Viviane Rangel, “Informe Sobre Políticas de Inteligencia Artificial y Derechos de Autor en América Latina”, 2023, Acceso Justo al Conocimiento, March 3, 2023, https://accesoalconocimiento.lat/2023/03/03/informe-sobre-politicas-de-inteligencia-artificial-y-derechos-de-autor-en-america-latina/.
25 Organisation for Economic Co-operation and Development (OECD), “OECD Policy Framework for Digital Security,” OECD, 2022, p. 31, https://read.oecd-ilibrary.org/science-and-technology/oecd-policy-framework-on-digital-security_a69df866-en#page31.
26 OECD, “Encouraging Vulnerability Treatment: Responsible Management, Handling, and Disclosure of Vulnerabilities,” OECD, February 3, 2021, p. 51, https://one.oecd.org/document/DSTI/CDEP/SDE(2020)3/FINAL/en/pdf.
