Seção Especial

Rotas cruzadas: governança da inteligência artificial e direitos humanos na América Latina

Uma governança da IA centrada no ser humano e implicações políticas latino-americanas


Atualmente estão em curso esforços para desenvolver quadros normativos para a governança da IA em diferentes níveis, incluindo a região latino-americana. Esses quadros normativos tendem a realçar que o “respeito e a proteção dos direitos humanos” devem estar no centro do desenvolvimento da IA. No entanto, o significado dessa expressão ainda precisa ser explorado. Este trabalho visa compreender as implicações políticas distintas que surgem para os Estados latino-americanos ao adotarem uma abordagem centrada no ser humano para a governança da IA.


inteligência artificial; governança de IA; América Latina; direitos humanos.
Imagem: Shutterstock

It is undeniable that artificial intelligence (AI)[1] has emerged as one of the disruptive technologies of this century with the potential to trigger significant transformations in welfare, wealth, and power (Dafoe 2018, 8–9), comparable to the agricultural or industrial revolutions (Karnofsky 2016). In the coming years, it is expected that AI will permeate nearly every aspect of our lives. As a result, a range of concerns will require careful attention, particularly those regarding AI governance.

Currently, there is a global race to establish the normative framework for AI development reflecting the interests of multiple stakeholders encompassing States and non-State actors such as academia, industry, and civil society. Most of the endeavors, whether from the public or the private sector, have focused on shaping ethical and technical principles (Robles Carrillo 2020, 10) as the basis for AI governance.

This approach poses a risk, as an excessive focus on ethical and technical principles tends to overlook legal implications of AI governance (Robles Carrillo 2020, 10ff). For instance, AI initiatives are often drawn from the philosophical discipline of ethics and, as a result, they show a tendency to overlap (and sometimes confuse) ethical principles with human rights (Jones 2023, 10). This is problematic since they are two well-defined fields that cannot substitute for the other (Jones 2023, 12).

AI initiatives are often drawn from the philosophical discipline of ethics and, as a result, they show a tendency to overlap (and sometimes confuse) ethical principles with human rights.

Most initiatives do not expressly refer to human rights at all. However, international efforts, such as the European Union’s (EU) Ethics Guidelines for Trustworthy AI (2019) and the Institute for Electrical and Electronics Engineers’ (IEEE) Ethically Aligned Design (2019), highlight the protection of human rights as a key component of AI governance.

At the regional level, the Organization of American States (OAS) has not been indifferent to these initiatives aiming to regulate the development and use of AI. In 2022, the IX Summit of the Americas adopted the Regional Agenda for Digital Transformation. In this document, the American States committed to promote “responsible and ethical development and use of artificial intelligence systems and other transformative technologies while protecting individual privacy and personal data and promoting equity and respect for human rights(OAS 2022, 12–13).  

In contrast with other initiatives, like the EU’s Guidelines, the American document does not provide a clear understanding of what “protection for human rights” stands for within the context of IA governance. Given that interpretation of legal provisions is one of the crucial aspects for international law enforcement, it is relevant to comprehend the implications of human rights protection for AI governance. 

Taking into account the unique features of the Inter-American System, especially within the context of the system of human rights, this work explores the significance of the term “protection of human rights” regarding AI governance. Does this phrase merely serve as a broad provision lacking a specific legal content, or does it entail distinct obligations within the framework of the Inter-American System of Human Rights, as argued here?

At the regional level, the protection of human rights is a longstanding tradition; as a result, a vibrant protection system is in place. The system encompasses a series of human rights treaties, such as the American Convention on Human Rights (ACHR) (1969) or the “Protocol of San Salvador” (1989), that impose specific obligations on States parties and provide rights for individuals. Furthermore, the system is overseen by two organs: the Inter-American Commission of Human Rights and the Inter-American Court of Human Rights (IACHR).

As the judicial organ of the system, the IACHR has the authority to interpret and apply the Inter-American human rights treaties (OAS 1969 art. 62). This competence raises significant questions, such as whether the Court, in exercising its functions, can establish the scope of AI governance (for American States) by interpreting States’ obligations within the human rights system; and, if so, to what extent? Moreover, can the Court establish obligations for private companies based on the business and human rights approach? 

Using doctrinal analysis, this work addresses these relevant questions for AI governance. The primary focus of this study is to outline policy implications for Latin American States, as they represent the majority of States parties to the ACHR[2]. Given their substantial representation within the ACRH framework, these States are more susceptible to become legally bound and be substantially affected by decisions arising from the IACHR. 

The next section briefly examines what AI governance entails and looks at the particularities of AI governance in Latin America. The third section explores the interplay between human rights and artificial intelligence, showing the significance of human rights approach for AI governance. Then, the fourth section describes how human rights are protected in Latin America and the implications of the IACHR’s role in shaping the system. The fifth section discusses which are the policy implications for Latin America of the Inter-American human rights system in the context of AI governance. Finally, this work concludes that Latin American States should adopt a human-centered approach when addressing AI governance challenges. While embracing this approach, States must remain aware of the distinctive obligations emanating from the Inter-American human rights system.

Artificial Intelligence Governance

AI governance proposes to understand how society should manage the transition towards advanced AI systems across different dimensions, including the political, ethical, and economic spheres (Dafoe 2018, 5). AI governance encompasses a variety of “tools, solutions, and levers that influence AI development and applications” (Butcher & Beridze 2019, 88), which includes establishing normative frameworks – whether ethical, technical or legislative – in the AI landscape.

AI governance is necessary because, despite AI technology’s moral neutrality (e.g. neither intrinsically good or evil), it introduces governance challenges that need to be addressed (Büthe et al. 2022, 1725). Governments should be able to manage the risks arising from AI usage while simultaneously promoting AI innovation (Graeme et al. 2022, 1823; Taeihagh 2021, 138). Among the concerns raised by AI are biases on data and algorithms employed for AI system training, potentially leading to unfair decision-making processes (Bello y Villarino & Vijeyarasa 2022, 198–99), job displacement due to automation, and the potential weaponization and malicious utilization of AI technologies.

In recent years, significant efforts have been made to address these challenges. Ongoing initiatives aim to establish normative frameworks to regulate AI’s development, deployment, and usage. Stakeholders, including governments, industry, academia, and civil society, are actively engaged in discussions looking to identify the most suitable approaches for governing AI. These proposals cover a wide spectrum ranging from non-binding norms such as principles, standards, and guidelines, to legally binding measures like laws and regulations crafted at both national and international levels (Büthe et al. 2022, 1725). However, these approaches vary significantly depending on the particular stakeholder involved (Butcher & Beridze 2019, 89).

As AI governance is a developing field, there are no widely agreed-upon classifications for models of AI governance. Most authors tend to describe the different approaches following the stakeholders involved in the initiative. This work follows a similar approach and thus identifies private and public sector initiatives. Private sector initiatives include those proposed by private companies and the proposals from NGOs, academia, and civil society, while public sector initiatives refer to those proposals made by governments and international organizations.

Private sector initiatives

Among the prominent strategies in AI governance, self-regulation by the private sector stands out. In this model, usually the same companies engaged in the development and deployment of AI technologies are the ones crafting and committing to uphold a set of specific guidelines or codes of conduct. 

These guidelines or codes span from general ideals for beneficial AI development to specific engineering standards for AI. Typically, they serve a dual purpose: establishing norms of conduct and results, and fostering the conditions to achieve those norms (Boddington 2020, 126). Notable examples of this kind of model are Google’s AI principles, Microsoft’s AI principles, IBM’s Everyday Ethics for Artificial Intelligence, and Principles for Trust and Transparency.

Other initiatives within the private sector include those developed by non-governmental and research organizations. These institutions also frequently propose a model of self-regulation by private companies. However, in this scenario, companies are required to adhere to guidelines or standards developed by these institutions, which usually draw upon ethical and technical principles. A notable example of this kind of initiative is the IEEE’s Ethically Aligned Design, which standardizes the creation of AI systems by integrating ethical considerations throughout the lifecycle of AI development, from design to deployment.

One of the advantages of the self-regulation model lies in the very ethical or technical principles that inform these guidelines or codes of conduct. These principles not only ensure risk mitigation in AI technology usage, but also offer flexibility, preventing hindrances to the development of these technologies. As a result, guidelines possess the capacity to adapt swiftly and undergo revisions as necessary (Taeihagh 2021, 145). However, this very characteristic raises concerns, as companies could potentially manipulate these principles to serve their own interests (Mantelero & Esposito 2021, 4–5; Yeung, Howes & Pogrebna 2020, 80).

Public sector initiatives

Public sector initiatives refer to proposals put forth by governments and international organizations. Governments face several challenges when attempting to establish regulatory frameworks for technology usage. In the case of AI, governments grapple with finding a balance between regulation and responsible development of AI technologies (Belli & Zingales 2022, 2). There is a belief that establishing legally binding regulatory frameworks will lead to limitations on AI development (Yeung, Howes & Pogrebna 2020, 79). However, due to AI’s pervasive nature, it is crucial to have appropriate regulatory frameworks in place, as not all organizations will act responsibly (Clarke 2019, 398).

In the face of these challenges, one of the strategies that has been employed involves the development of national AI strategies. They could be conceptualized as “public policies that can encompass guidelines and objectives that governments set out to plan, implement, and assess the use of AI in various domains of government, society, and the private sector” (Scrollini, Cervantes & Mariscal 2021, 6–7, my translation). These documents allow States to outline the fundamental goals regarding AI development. As a result, they identify the priority areas where development efforts should concentrate, and they establish mechanisms to achieve those goals. Additionally, these documents recognize those stakeholders considered crucial for AI development (Djeffal, Siewert & Wurster 2022, 1806).  

The development of national strategies is influenced by multiple factors since States must balance the interests of various stakeholders. This leads national strategies to typically aim at establishing codes of conduct for different stakeholders in the AI environment, rather than imposing binding regulations on diverse actors. As a result, these strategies are often inspired by ethical principles. However, the development of these documents is crucial as they can serve as the foundation for the development of more complex and potentially legally binding regulations. Examples of this kind of strategy include those prepared by the United Kingdom, Germany, and the United States.

International organizations have not been oblivious to these discussions. There are several efforts underway aimed at addressing issues related to AI governance. These efforts vary significantly depending on the structure of the international organization and the stakeholders involved in the discussions. So far, the most comprehensive efforts have occurred within the context of the European Union, which has been taking actions, since at least 2017, to develop a regulatory framework for governing the use and development of AI. Some notable documents in this regard include the proposal for the EU AI Act (2021), which aims to establish a comprehensive regulatory framework for AI in the European Union, and the High-Level Expert Group’s document Ethics Guidelines for Trustworthy AI (2019) (see, e.g., Nikolinakos 2023).

On the global level, the United Nations has sponsored various dialogues aimed at examining the balance between the potential risks of AI utilization and its development. Within the United Nations system, there are underway multiple initiatives in this regard (see, e.g., Butcher & Beridze 2019, 92–93). Perhaps one of the most prominent examples in this regard is UNESCO’s Recommendation on the Ethics of Artificial Intelligence (2021), which aims “to provide a universal framework of values, principles, and actions to guide States in the formulation of their legislation, policies or other instruments regarding AI, consistent with international law” (UNESCO 2022, 14).

Overall, a distinct characteristic shared by documents produced at both national and international levels is that the protection and respect for human rights are often regarded as core values for AI governance. However, in most cases, these documents do not explain what the protection or respect for human rights entails within AI governance. They refer neither to specific rights nor to specific human rights treaties. A notable exception is the EU’s Act, which expressly refers to the European human rights system.

Artificial Intelligence Governance in Latin America

At the regional level, AI deployment and usage in Latin America are on the rise. It is commonly employed by both the private and the public sector (Belli & Zingales 2022, 1–2). However, as noted by Sanchez-Pi et al. (2021, 3), AI development accounts for only 0.5% of private investment in the region. As a result, while there is increasing awareness and eagerness to establish appropriate regulatory frameworks, AI governance is in its early stages (Belli & Zingales 2022, 2). 

Comprehensive efforts to establish a regional framework for AI governance have yet to materialize. As a result, most State efforts are focused on developing and implementing national frameworks or strategies. Nonetheless, these efforts fall short as Latin American AI strategies are neither consolidated nor sustainable in the long term (Scrollini, Cervantes & Mariscal 2021, 3). 

Studies addressing AI governance in the region highlight that the approach of Latin American States to AI is not uniform across the region. While States have developed documents related to AI, they are not typically formulated with the purpose of promoting specific actions concerning the development and implementation of AI.

Studies addressing AI governance in the region highlight that the approach of Latin American States to AI is not uniform across the region. While States have developed documents related to AI, they are not typically formulated with the purpose of promoting specific actions concerning the development and implementation of AI. Instead, these documents often take a general stance, outlining a desired point to reach in AI matters, and primarily addressing local needs (Scrollini, Cervantes & Mariscal 2021, 10; Prudencio Ruiz 2021, 4). Consequently, Latin America lacks a cohesive AI framework and exhibits varying levels of development and implementation of AI strategies, contingent upon the specific State under analysis (Prudencio Ruiz 2021, 4, 6–7; Belli & Zingales 2022, 3).

In this context, it is interesting to highlight the findings emerging from the recent Latin American Artificial Intelligence Index (2023), which seeks to provide an overview of the state of AI in the region. Regarding AI governance, the index reveals a significant gap between countries that have achieved the highest performance scores (Brazil, Chile, and Argentina) and those at the opposite end of the spectrum (Bolivia, Ecuador, Panama, and Paraguay), which have received zero scores in certain dimensions analyzed (CENIA 2023, 116–17).

Although a regional framework does not currently exist, Latin American States' AI strategies share two common elements: firstly, they aim to devise mechanisms to address local issues; secondly, they are developed with the objective of promoting the enjoyment of the benefits associated with AI utilization and its advancement (Prudencio Ruiz 2021, 10). In this context, it is often pointed out that the formulation of AI strategies presents a unique opportunity to design approaches that reduce the inequalities present in Latin America (CENIA 2023, 100–101; Marzetti 2021).

Another noteworthy characteristic of these regional processes is the parallel development of AI strategies and regulatory frameworks concerning personal data protection. Examples include those developed in Brazil and Chile (see, e.g., Castaño 2020, 14ff).

At the level of regional organizations, specific regulatory frameworks are also absent. Nonetheless, the OAS has produced certain documents in which it underscores the importance of regulating the promotion of artificial intelligence. For instance, the IX Summit of the Americas adopted the Regional Agenda for Digital Transformation. Although it is not a dedicated document for AI governance, it articulates a perspective on how the American States should engage with AI technologies, emphasizing that equity and human rights should be at the core of the development of AI technologies (OAS 2022, 12–13).

Like the international context, Latin American strategies and regional documents concerning AI emphasize that “protection or respect for human rights” are core values when defining AI policies (CENIA 2023, 172). However, like most existing documents, they lack definitions explaining what “protection of human rights” entails within IA governance.

Nevertheless, given the existence of a human rights protection system to which Latin American States are signatories, could this expression carry a distinctive significance? This question will be explored in the forthcoming sections.

Interaction between Human Rights and Artificil Intelligence

To truly grasp the significance of including human rights in AI governance discussions in Latin America, it is crucial to examine how the human rights framework and AI interact in a broader context. As a result, this section aims to explore the relationship between human rights and AI, providing insights into their interconnection.

In the past few years, a growing number of scholars has engaged in discussing the relevance and implications of human rights for AI governance (see, e.g., Gordon 2023; Jones 2023; Bello y Villarino & Vijeyarasa 2022; Greiman 2021; Yeung, Howes & Pogrebna 2020; Risse 2019; Raso et al. 2018). With the widespread adoption of AI technologies and their pervasiveness in everyday life, the likelihood of their impact on human rights increases significantly. This is particularly significant when considering the growing utilization of these technologies in decision-making processes. In such cases, they can lead to significant life-changing consequences for the persons involved in those processes (Yeung, Howes & Pogrebna 2020, 78). Examples include loan applications or a parole-related risk assessment which were made using biased algorithms and datasets (see, e.g., Raso et al. 2018, 18ff).

Studies on human rights interaction with AI highlight that human rights must be considered when addressing AI governance. Human rights framework can provide the necessary tools to address the challenges posed by AI design, development, and deployment. As stated by Latonero (2018, 1) “[i]nternational human rights are a powerful tool for identifying, preventing, and redressing an important class of risks and harms”.

There are several advantages for using a human rights-centered approach. One of them is that international human rights law provides for a framework aimed at the protection of the human dignity of individuals by recognizing their rights while establishing accountability mechanisms to oversee States duties. As a result, it provides stakeholders with a guidance for upholding the inherent human dignity of each person regardless of where they are situated (Yeung, Howes & Pogrebna 2020, 81). Moreover, this framework rests on a broad global consensus and establishes a universal set of norms and commitments (Greiman 2021, 54; Donahoe & MacDuffee Metzger 2019, 119) articulated on international treaties. 

International human rights law is constantly evolving. The normative framework is enriched through the adoption of new international documents on human rights, the practices of supervisory bodies, and both national and international jurisprudence (Raso et al. 2018, 8). As a result, these elements collectively contribute to clarify and expand the scope of human rights.

It is necessary to highlight that international human rights law is built upon specific State obligations to prevent and safeguard human rights. The normative framework also rests on distinct mechanisms to assess compliance with these obligations. When assessing compliance with the human rights framework, State conduct is evaluated. The State needs to make sure that there are in place suitable mechanisms, like the adoption of regulatory frameworks, to prevent human rights violations not only by its agents but also by third parties, such as corporations. 

In the field of AI governance, these obligations are particularly relevant. On the one hand, the State is accountable for violations stemming from the use of AI technologies in its processes – for instance, algorithm-based immigration decisions. On the other hand, the State is responsible for ensuring that private activities do not harm individuals (see McGregor, Murray & Ng 2019, 327–29).  

However, over the past decades, the international human rights framework has expanded to encompass certain expectations for businesses concerning human rights that are independent of States obligations (UN HRC 2011, 13). Documents like the United Nations Guiding Principles on Business and Human Rights (2011) reflect the businesses commitment in terms of human rights protection. For instance, they provide for the expectation that businesses should avoid causing or contributing to adverse human rights and to prevent or mitigate adverse human rights impact. Moreover, businesses should also redress any adverse human rights impact.

In the context of AI governance, the companies that are designing, developing, and deploying AI do not operate in a legal vacuum. Therefore, these commitments arising from the international human rights law are relevant as they establish the normative framework against which their actions must be measured.

In the context of AI governance, the companies that are designing, developing, and deploying AI do not operate in a legal vacuum. Therefore, these commitments arising from the international human rights law are relevant as they establish the normative framework against which their actions must be measured.

Is this interaction characterized by specific attributes within the Latin American context? The next section provides a brief background on the human rights protection system currently in place in Latin America. 

The Protection of Human Rights in Latin America

A brief background of the human rights protection system in place in Latin America is necessary to comprehend distinctive characteristics of this normative framework in the region. Consequently, it helps in understanding the policy implications for Latin American States of adopting a human rights approach to AI governance. 

The Inter-American human rights system could be described as a network of principles, norms, and institutions crafted to protect the rights of individuals. Given its subsidiary nature, the primary obligation to respect and guarantee human rights rests with the States (Hennebel & Tigroudja 2022, 22). However, the international bodies established by the American Convention on Human Rights (ACHR), the Inter-American Commission on Human Rights, and the Inter-American Court of Human Rights oversee State behavior (Pasqualucci 2013, 3).

The Inter-American human rights system has a complex normative framework and offers varying levels of protection based on the applicable instruments for each specific State. Since the majority of Latin-American States are party to the ACHR, the system tends to provide the highest standards of protection. Consequently, the IACHR is empowered to fulfill its functions, and cases of human rights violations can be addressed by this judicial institution.

Drawing from the normative framework, Latin American States have different duties. The first one is to respect the human rights recognized in the treaties and to ensure their full and free exercise. This duty has a negative aspect (“to respect”) which requires that States or their organs refrain or abstain from interfering with the exercise of rights; the positive aspect (“to ensure”) requires that States take the necessary action to protect the rights allowing individuals to be able to exercise their rights (Antkowiak & Gonza 2017, 19; Medina 2017, 17–19). This obligation also requires that States take all the necessary domestic measures to guarantee those rights, encompassing not only the adoption of laws and regulations but also adopting specific behaviors depending on the right in question (Hennebel & Tigroudja 2022, 28).

Various mechanisms are in place within the system to supervise and verify the fulfillment of State obligations. These mechanisms encompass a spectrum of activities, ranging from the submission by the Inter-American Commission on Human Rights of reports on a State’s human rights situation to the delivery of binding decisions by the Inter-American Human Rights Court in specific cases where a State’s violation of human rights has been established.

The IACHR is the sole judicial organ of the system. As such, it is the final interpreter of the ACHR (Dulitzky 2015, 69). This means it can define the scope of human rights recognized in various Inter-American treaties and, thereby, the extent of States' obligations concerning these rights. As a result, the role the Court plays in shaping the system is crucial. 

Throughout its history, the Court has utilized interpretive tools to expand the scope of its jurisdiction and the rights contained within the system's instruments (Lixinski 2010, 586). As a result, the Court engages in an “evolutive interpretation” of human rights treaties (Neuman 2008, 107). 

One of these tools is the conventional control doctrine developed by the IACHR. It is defined as an “instrument for applying international law” (IACHR 2013), allowing domestic judges to directly apply international norms and standards of interpretation. The IACHR understands that domestic judges should also follow the IACHR’s interpretation of the ACHR (Contesse 2018, 1172). This is particularly contentious because States not parties to the cases have no legal obligations to follow the IACHR’s interpretation (Dulitzky 2015, 70).

As a result, by resorting to this tool, IACHR is able to shape the Inter-American system of human rights because the extent of State obligations and the scope of human rights are determined by the Court’s understanding of them. Domestic courts could follow these “precedents” for judicial economic reasons and to avoid international State responsibility. In this context, the IACHR could have a potential impact on AI governance.

The Inter-American Human Rights Framework and its Policy implications for AI Governance in Latin America

Most of the documents related to AI governance in Latin America, along with studies on these documents, emphasize the importance of defining context-specific solutions tailored to the needs of States of the region (see, e.g., CENIA 2023, 117; Levy Daniel 2023). Therefore, when defining AI governance policies in Latin America, it is necessary to consider the peculiarities of the Inter-American human rights system, especially those related to the role of the IACHR in shaping the system.

As noted in a previous section, international human rights law can provide a useful normative framework for AI governance. In Latin America, States are bound by a robust human rights protection system, which is largely shaped by the IACHR’s interpretations of the human rights instruments within the system. Thus, the IACHR specifies the content of State obligations and the scope of rights recognized in these instruments.

By defining the scope of those rights and obligations, the Court sets a standard of conduct for the States bound by those treaties. Although IACHR's role as the ultimate interpreter of the Convention has been contested, Latin American States are still required to follow the standard so defined to avoid international responsibility. 

As a result, this work argues that some policy implications follow from this distinctive characteristic of the system. AI governance does not exist in a legal vacuum; it must be understood within the international normative framework to which States are bound. As AI usage can potentially impact human rights, Latin American States should consider a human rights-centered approach when addressing AI governance, because international human rights law provides powerful tools to address the challenges posed by AI.

In this regard, when Latin American States commit to respecting and protecting human rights in the context of AI governance, they are assuming specific content-related obligations. They are referring to the set of international obligations they have concerning human rights, both at the universal and regional levels. At the regional level, these obligations fall within the specific context of the Inter-American human rights protection system, which entails accepting the IACHR's interpretations of treaty content.

Although cases directly related to the establishment of regulatory frameworks for AI have not yet been brought before the IACHR, its decisions in other cases could influence their development. As mentioned earlier, AI pervasiveness in everyday life can potentially impact human rights. Hence, when the IACHR addresses issues such as non-discrimination, procedural guarantees, the right to privacy, protection of personal data, responsibility of businesses regarding human rights, it can establish a standard of conduct that Latin American States should consider when designing AI policies to avoid international responsibility.

Although cases directly related to the establishment of regulatory frameworks for AI have not yet been brought before the IACHR, its decisions in other cases could influence their development. As mentioned earlier, AI pervasiveness in everyday life can potentially impact human rights.

The design of AI policies that adopt a person-centered approach will necessarily be linked to the protection provided by the human rights system. Therefore, for their effective implementation, multi-sectoral dialogue will be required to design policies that align with the realities of the region. 

The human rights protection system does not have to become a limiting factor for the development of AI at the regional level. It can serve as the foundation for constructing normative frameworks that place the individual at the center of AI developments. In this sense, regular assessments of AI governance strategies, informed by ongoing discussions within the human rights framework, will enable Latin American States to effectively navigate the intricate landscape of AI and human rights.

Latin American States obligations in the context of AI governance are shaped by the region’s distinctive human rights protection system. This framework calls for an AI governance approach that is designed and firmly rooted in the principles of human rights protection. By adopting this approach, Latin American States, along with other stakeholders, can work towards achieving a responsible and ethical deployment of AI technologies that also respects human rights.


The international human rights framework serves a powerful mechanism that should be considered when dealing with AI governance. It provides tools for effectively tackling the plethora of concerns arising from the widespread presence of AI in daily life.

Within Latin America, the landscape of AI governance requires bespoke solutions to address region-specific challenges. When formulating dedicated AI normative frameworks, Latin American States should adopt a human-centered approach, positioning individuals at the core of AI advancements. However, while embracing this approach, States must remain aware of the distinctive obligations emanating from the Inter-American human rights system.

The international human rights framework serves a powerful mechanism that should be considered when dealing with AI governance. It provides tools for effectively tackling the plethora of concerns arising from the widespread presence of AI in daily life.

A relevant actor in shaping State obligations, thereby influencing companies engaged in developing AI technologies, is the IACHR. Through the interpretation of human rights treaties, this institution has the potential to delineate specific responsibilities and obligations for States. Nonetheless, this dynamic should not be construed in a negative light; rather it underscores the need for ongoing consultations amongst stakeholders involved in AI design, development, and deployment.

Rather than acting as a constraining force, human rights stand as a robust and adaptable framework that provides effective tools for addressing the inherent risks and challenges that arise with the advent of AI. 


[1] There is not a commonly agreed conceptualization of artificial intelligence. In this work artificial intelligence is understood “as a general-purpose technology, for automating and improving the accuracy, speed and/or scale of machine decision-making, pattern-recognition and prediction in complex or large environments, with the aim of substituting for, or improving upon, human performance in specific tasks” (Maas 2019, 2).

[2] Only 24 out of 35 American States are parties to the ACHR. They are from the Latin American and the Caribbean region. Neither Canada nor the United States of America are parties to the Convention. 


AI HLEG. 2019. Ethics Guidelines for Trustworthy  AI. Brussels: European Commission Report.

Antkowiak, Thomas M. & Alejandra Gonza. 2017. The American Convention on Human Rights. Essential Rights. New York: Oxford University Press.

Belli, Luca & Nicolo Zingales. 2022. “Data Protection and Artificial Intelligence Inequalities and Regulations in Latin America”. Computer Law & Security Review 47 (105761): 1–3.

Bello y Villarino, Miguel & Ramona Vijeyarasa. 2022. “International Human Rights, Artificial Intelligence, and the Challenge for the Pondering State: Time to Regulate?” Nordic Journal of Human Rights 40 (1): 194–2145.

Markus D. Dubber, Frank Pasquale & Sunit Das. 2020. “Normative Modes”. The Oxford Handbook of Ethics of AI.

Butcher, James & Irakli Beridze. 2019. “What is the State of Artificial Intelligence Governance Globally?” The RUSI Journal 164 (5–6): 88–96.

Büthe, Tim, Christian Djeffal, Christoph Lütge, Sabine Maasen & Nora von Ingersleben-Seip. 2022. “Governing AI – Attempting to Herd Cats? Introduction to the Special Issue on the Governance of Artificial Intelligence”. Journal of European Public Policy 29 (11): 1721–1752.

Castaño, Daniel. 2020. “La gobernanza de la Inteligencia Artificial en América Latina: entre la regulación estatal, la privacidad y la ética digital”.

CENIA. 2023. “Índice Latinoamericano de Inteligencia Artificial”. Centro Latinoamericano de Inteligencia Artificial.

Clarke, Roger. 2019. “Regulatory Alternatives for AI”. Computer Law & Security Review 35 (4): 398–409.

Contesse, Jorge. 2017. “The International Authority of the Inter-American Court of Human Rights: a Critique of the Conventionality Control Doctrine”. International Journal of Human Rights 22 (9): 1168–1191.

Dafoe, Allan. 2018. “AI Governance: a Research Agenda”. Oxford: Centre for the Governance of AI Future of Humanity Institute. University of Oxford.

Djeffal, Christian, Markus B. Siewert & Stefan Wurster. 2022. “Role of the State and Responsibility in Governing Artificial Intelligence: a Comparative Analysis of AI Strategies”. Journal of European Public Policy 29 (11): 1799–1821.

Donahoe, Eileen & Megan MacDuffee Metzger. 2019. “Artificial Intelligence and Human Rights”. Journal of Democracy 30 (2): 115–26.

Dulitzky, Ariel E. 2015. “An Inter-American Constitutional Court? The Invention of the Conventionality Control by the Inter-American Court of Human Rights”. Texas International Law Journal 50 (1): 45–93.

Gordon, John-Stewart. 2023. The Impact of Artificial Intelligence on Human Rights Legislation. A Plea for an AI Convention. Switzerland: Palgrave Macmillan.

Graeme, Auld, Ashley Casovan, Amanda Clarke & Benjamin Faveri. 2022. “Governing AI through Ethical Standards: Learning from the Experiences of Other Private Governance Initiatives”. Journal of European Public Policy 29 (11): 1822–1844.

Greiman, VA. 2021. “Human Rights and Artificial Intelligence: a Universal Challenge”. Journal of Information Warfare 20 (1): 50–62.

Hennebel, Ludovic & Hélène Tigroudja. 2022. The American Convention on Human Rights. A Commentary. Oxford: Oxford University Press.

Institute for Electrical and Electronics Engineers. 2019. “Ethically Aligned Design: a Vision for Prioritizing Human Well-Being with Autonomous and Intelligent Systems. First Edition.” IEEE.

Inter-American Court of Human Rights. Gelman v. Uruguay, Judgment of 20 March 2013 (Monitoring Compliance with Judgement), para. 60.

Jones, Kate. 2023. AI Governance and Human Rights. Resetting the Relationship. London: Chatham House.

Karnofsky, Holden. 2016. “Potential Risks from Advanced Artificial Intelligence: the Philanthropic Opportunity”. Open Philanthropy (blog). 6 May 2016.

Latonero, Mark. 2018. “Governing Artificial Intelligence:  Upholding Human Rights & Dignity”. Data & Society.

Levy Daniel, Maia. 2023. “AI Regulation in Latin America Requires a Thoughtful Process”. Tech Policy Press (blog). 07 2023.,tools%20need%20to%20be%20vetted.

Lixinski, Lucas. 2010. “Treaty Interpretation by the Inter-American Court of Human Rights: Expansionism at the Service of the Unity of International Law”. European Journal of International Law 21 (3): 585–604.

Maas, Matthijs M. 2019. “International Law Does not Compute: Artificial Intelligence and the Development, Displacement or Destruction of the Global Legal Order”. Melbourne Journal of International Law 20 (1): 29–57.

Mantelero, Alessandro & Maria Samantha Esposito. 2021. “An Evidence-Based Methodology for Human Rights Impact Assessment (HRIA) in the Development of AI Data-Intensive Systems”. Computer Law & Security Review 41 (105561): 1–35.

Marzetti, Maximiliano. 2021. The Politics of Technology in Latin America. Data Protection, Homeland Security and the Labor Market. Oxon and New York: Routledge.

McGregor, Lorna, Daragh Murray & Vivian Ng. 2019. “International Human Rights Law as a Framework for Algorithmic Accountability”. International Law & Comparative Law Quarterly 68 (2): 309–343.

Medina, Cecilia. 2017. The American Convention on Human Rights.  Crucial Rights and their Theory and Practice. 2nd ed. Cambridge, Antwerp, Portland: Intersentia.

Neuman, Gerald L. 2008. “Import, Export, and Regional Consent in the Inter-American Court of Human Rights”. European Journal of International Law 19 (1): 101–123.

Nikolinakos, Nikos Th. 2023. “EU Policy and Legal Framework for Artificial Intelligence, Robotics and Related Technologies - The AI Act.” Law, Governance and Technology Series 53.

Organization of American States. 1969. American Convention on Human Rights.

Organization of American States. IX Summit of the American States. 2022. “Mandates Adopted at the IX Summit of the Americas”. Los Angeles. 

Pasqualucci, Jo M. 2013. The Practice and Procedure of the Inter-American Court of Human Rights. 2nd ed. New York: Cambridge University Press.

Prudencio Ruiz, Karla. 2021. “Estrategias de Inteligencia Artificial en América Latina. Conclusiones a Nivel Regional”. Empatía

Raso, Filippo, Hannah Hilligoss, Vivek Krishnamurthy, Christopher Bavitz & Levin Kim. 2018. “Artificial Intelligence and Human Rights: Opportunities & Risks”. Boston: Berkman Klein Center for Internet & Society at Harvard University.

Risse, Mathias. 2019. “Human Rights and Artificial Intelligence: an Urgently Needed Agenda”. Human Rights Quarterly 41 (1): 1–16.

Robles Carrillo, Margarita. 2020. “La gobernanza de la Inteligencia Artificial: contexto y parámetros generales”. Revista Electrónica de Estudios Internacionales, no. 39: 1–27. 10.17103/reei.39.07.

Sanchez-Pi, Nayat, Luis Martí, Ana Cristina Bicharra García, Ricardo Baeza Yates, Marley Vellasco & Carlos Artemio Coello Coello. 2021. “A Roadmap for AI in Latin America.” Side event AI in Latin America of the Global Partnership for AI (GPAI) Paris Summit, GPAI Paris Center of Expertise, Nov 2021, Paris, France.

Scrollini, Fabrizio, María Esther Cervantes & Judith Mariscal. 2021. “En busca de rumbo: el estado de las políticas de Inteligencia Artificial en América Latina”. Empatía.

Taeihagh, Azar. 2021. “Governance of Artificial Intelligence”. Policy and Society 40 (2): 137–57.

United Nations Human Rights Council. 2011. “Guiding Principles on Businesses and Human Rights”.

United Nations Educational, Scientific and Cultural Organization. 2022. “Recommendation on the Ethics of Artificial Intelligence”.

Yeung, Karen, Andrew Howes & Ganna Pogrebna. 2020. “AI Governance by Human Rights-Centered Design, Deliberation and Oversight: an End to Ethics Washing”. The Oxford Handbook of Ethics of AI: 77–106. 

Recebido: 25 de agosto de 2023

Aceito para publicação: 11 de setembro de 2023

Copyright © 2023 CEBRI-Revista. Este é um artigo em acesso aberto distribuído nos termos da Licença de Atribuição Creative Commons que permite o uso irrestrito, a distribuição e reprodução em qualquer meio desde que o artigo original seja devidamente citado.