In the last three years, and especially since the widespread adoption of large language models (LLMs) such as ChatGPT, LaMDA, and SORA, artificial intelligence has become central to public discourse. Its impact on politics, economics, work, culture, the arts, health, and education, among other areas, is already a reality in various countries and societies, and this impact is expected to become increasingly profound and accelerate in the near future.

In this context, debates about the governance of these socio-technical systems, and about how they affect democratic quality, the field of public information, the worlds of work, social institutions and subjective configurations, occupy governments, universities, international organizations, technology corporations and representatives of civil society.

The technical, social, and political opacity that characterizes these systems, their interactive complexity, and their partially autonomous behavior present a real challenge and demand a conceptualization effort that has only just begun but is undoubtedly urgent. What exactly are these metatechnologies? At what scales do they operate? What are the benefits and dangers of their widespread deployment? What role can states, scientific and pedagogical knowledge, and social organizations and movements play in addressing them?

The overall objective of this Working Group is to build a regional agenda and a robust set of tools to address these questions from Latin America, with the purpose of enriching the scientific, political, legal, artistic and philosophical efforts that are being made to grasp the dimension of the transformation that is coming.

Among the Group's main antecedents are the Higher Diploma in Technology, Subjectivity, and Politics and the Higher Diploma in Artificial Intelligence and Social Sciences: Theoretical and Epistemological Challenges, which are co-directed at CLACSO by one of the Working Group's coordinators (Flavia Costa), and which held their fifth and second editions, respectively, in 2025. The vast majority of the instructors from these Higher Diplomas are also members of this Working Group as researchers. Scientific meetings and thematic networks have also been established (such as the International Colloquium on Philosophy of Technology, now in its 14th edition, and the LAVITS Network: Latin American Network for Studies on Surveillance, Technology, and Society, the KHIPU Network, among many other regional initiatives).

Our region participates in this global discussion with its own unique challenges. These include:

1) The need to develop regulations capable of protecting fundamental and diverse rights such as privacy, copyright, and labor rights, as well as the legal force necessary for their effective enforcement. In this regard, it is worth mentioning the Brazilian Artificial Intelligence Strategy (2021); Chile's National Artificial Intelligence Policy (2021, updated in 2024); Argentina's Recommendations for Trustworthy AI (2023); and Uruguay's National Artificial Intelligence Strategy (2024), as well as the Montevideo Declaration (2023). Within this framework, it is expected that the region will experience a period of intense regulatory activity in the coming years.

2) Currently based in Chile, the Latam GPT project—a collaborative, large-language model powered by scientifically verified datasets from over twenty countries in the region—continues the discussion on whether it is possible to develop artificial intelligence beyond that produced in global centers. In this sense, it is a project that proposes traceability, open data and open source, transparency, and representation of communities overlooked by datasets from the Global North, and it serves as a key case study for understanding what, for example, "technological sovereignty" means today. Several members of the Working Group collaborate on Latam GPT.

3) The term "artificial intelligence" and the metaphor of the "cloud" associated with the digital ecosystem obscure the fact that these technologies operate using highly tangible minerals and energy resources. Countries in our region are entering the AI ​​ecosystem within a new form of colonialism, as "sacrifice zones." This manifests itself in the struggle over environmental sustainability and in the conflicts that various communities are experiencing with the installation of Big Tech data centers.

4) There is a growing number of Latin American art projects working with Generative Adversarial Networks (GANs) and LLMs. While some replicate the aesthetic and political logics of hegemonic commercial art, we will focus on critical Latin American technopoetics, which adopt a decolonial perspective. We will examine the development of pieces and practices that resist at festivals such as Toda la teoría del Universo (Chile), La Colmena (Mexico), among others.

In March 2023, a group of experts from our region, meeting in Montevideo, highlighted "the productive potential of artificial intelligence systems, as well as the risks associated with their unreflective growth." They also pointed out the need to "develop criteria and standards that allow these technologies to be clearly and transparently assessed according to their risks, in order to advance public policies that protect the common good without hindering the benefits of technological development." This is due to the enormous power of AI to accelerate production processes and decision-making; because the introduction of AI impacts the world of work and education, forcing the rewriting of rules for entire industries; because of its capacity to instantly create content and news that may be false or erroneous; and because of its ability to generate situations in which existing regulations are no longer adequate to address the problems facing society, thus creating what are known as regulatory gaps.

In this context, this Working Group aims to provide tools to various technical and political teams—public policy developers, university authorities, those responsible for technology integration programs, teachers, researchers, and students from different disciplinary fields—to understand and organize existing information on initiatives related to Artificial Intelligence policies and regulations being developed in the region, in relation to other international reference initiatives.

It also seeks to deepen, from an interdisciplinary perspective, the development of an original and productive perspective that allows us to identify the impacts of these new technologies on democratic quality, citizen participation, education, and even the hierarchies between different fields of knowledge and disciplines, which are being profoundly affected by the ongoing transformation.

Finally, we aim to highlight the state of the latest scientific, artistic, and political debates on the issue, so that those who are part of teams responsible for implementing AI-related policies can identify the fundamental aspects that must be taken into account to carry out initiatives for the development, adoption, monitoring, and risk mitigation of a trustworthy, democratic, inclusive, and safe AI.

As a field of research, Artificial Intelligence is over 70 years old: its implementation coincides with World War II and the origin of the term requires going back to the 1950s, when in the United States they began to investigate and develop systems capable of replicating or emulating certain human behaviors considered intelligent: memorizing, computing, learning from their own results, projecting trends, making decisions.

However, it has been in recent years, and particularly since the end of 2022, with the massive deployment of chatGPT 3.5 and other Large Language Models (LLMs), that the question of what exactly these AI metatechnologies are and how to estimate their disruptive potential for the worlds of work, knowledge production, democratic ways of life, and subjectivities has become prevalent among governments, educational and research institutions, international bodies, and civil society organizations.

If one of the defining characteristics of our current moment is our intensive immersion in these complex socio-technical systems, then one of the urgent tasks is to address this question at both regional and global levels and formulate feasible answers. This is especially important considering that these are proprietary developments of large corporations within a platform-based context and amidst a multidimensional geopolitical struggle encompassing infrastructure, growing energy demand, the collection of behavioral and biometric data from entire populations, and content control.

The first contribution of the social sciences to this issue is to remind us that it is not a matter of deciding the future of "the species" in relation to advanced technologies, but rather of planning and guiding the development of these systems within societies marked by conflicts of interest, inequalities (of class, race, gender, geography, etc.), and ethical, linguistic, and cultural diversity. It is reasonable to predict that if this task is not undertaken soon, the uneven development of these technologies will further deepen the ongoing processes of stratification and inequality.

From a geopolitical perspective, AI is now a strategic battleground where global powers and technology corporations compete for control of information flows and cultural influence. In this “algorithmic governance” (Rouvroy and Berns 2016), algorithms not only regulate access to information but also generate narratives, shape perceptions, monitor communications, and modulate decisions and identities, wielding subtle yet profound power over society. This raises serious questions about digital sovereignty, privacy, and the balance of power on the international stage.

Emerging new AIs also impact subjectivities, as interactions with intelligent systems and algorithm-mediated digital environments influence how people perceive themselves and their surroundings.

For the social sciences, it has become imperative to address these issues if they wish to fully participate in a conversation that intimately concerns them in terms of their object of study—social life—but which, in practice, marginalizes them as experts in their field. Among other efforts, this scenario demands a revision of some of our theoretical and epistemological frameworks: distancing ourselves from perspectives that identify the technological dimension as merely instrumental, given the evidence that this digital world in and with which we interact is not just another tool, but a true environment, an ecosystem. Likewise, to the extent that these technologies are applied to decision-making in different spheres (public and private, academic, governmental, and business) and decisively influence the practices and structures of society; to the extent that they also create new spaces and habits for the production, distribution, and consumption of goods, services, and meanings, the social sciences must strive not only to be users or implementers of AI, but also to participate in the development and auditing of these systems.

In order to intervene critically and with a decolonial agenda in the digital and AI ecosystem, we propose mapping this complex and multi-layered territory by fully embracing the "materialist turn." Moving beyond the supposed "dematerialization" associated with digital technologies and instead questioning their materiality (submarine cables, cloud infrastructure, servers, satellites, rare earth mining) implies focusing on the geopolitical dynamics of the struggles over the various forms of extractivism that comprise it (from rare minerals to data), as well as mapping the political economy of the ecosystem of data, algorithms, and platforms. And along with this, it means accounting for the remnants, the traces, the "sacrifice zones" of transnational informational capitalism, where entire geographic regions are permanently subjected to environmental damage and a lack of investment.

The work agenda also includes the question of the place of mass surveillance and the remote activation of behavior within these digital ecosystems (Zuboff 2025). It also incorporates, as part of critical theory, the question of the inherent accidents of these highly complex technologies with which we coexist. The potential AI incidents that affect the quality of democratic life are numerous: from mass surveillance to deepfakes, from the proliferation of biases and the underrepresentation of minorities to the malicious manipulation of behavior, social classification, and epistemic inequalities. How do we collect and identify AI incidents in a timely manner? How do we study them? How do we anticipate them and "catch the error" to minimize their impact? This is not a task solely for computer scientists, but for interdisciplinary teams specializing in "artificial society." Finally, we ask ourselves how different social, cultural, and artistic collectives and movements are organizing to intercept and divert the neocolonialist and deeply authoritarian lines of force that are pressing to control the communicational, logistical, emotional, economic, and libidinal flows of our region.

It is for all these reasons that, in this GT, we propose to explore and expand four lines of work that we have identified as challenges of digital + AI ecosystems:

1. The epistemological challenges of this new stage of digitization + generative AI. The question of scales (beyond and beyond the individual-society relationship) and the Data-Algorithms-Platforms scheme as a model to analyze hypermediation processes (Rodríguez 2025).

2. The challenges of governance. On the one hand, the question of the accidents, risks, and dangers of digital ecosystems + AI. On the other, the threats to the information order and the threats to democracy (Zuboff 2025, Coeckelbergh 2025).

3. The impacts on ways of life and modes of subjectivation. The onto-technopolitical drift of this moment: are we heading towards a post/transhumanist model of humanity, or are we already there? How do we study and, eventually, contribute to reversing the processes of cognitive delegation, cognitive sedentarism (Sigman 2025), and other subjective impacts linked to transformations of attention (Heyles 2010)?

4) The resistances: critical epistemologies (feminisms, decolonial theories, alternative cosmopolitics), critical technopoetics and other practices of deviation, escape, profanation of the hegemonic digital device.