What bodies, what resources, what territories make the existence of artificial intelligence possible? This is the question that gives rise to this fanzine, an illustrated work that narrates the less visible yet utterly essential side of AI development. For decades, artificial intelligence has been presented as something ethereal, neutral, dematerialized. We have been led to believe that AI "thinks," "reasons," "feels"—that it exists on a superior plane, floating in clean and intelligent digital clouds. But this story is incomplete. What is not told and what we need to tell is that AI, data, technology is profoundly material. Behind every automated system, every recommendation, every algorithm, there are extracted minerals, diverted rivers, perforated mountains, and displaced forests.

AI needs lithium, copper, silicon, water, energy. To function, it requires data centers that consume more electricity than entire countries. It needs a colossal physical infrastructure, sustained by territories that are often colonized, exploited, or precarized. This infrastructure has place and body.But it is not only resources. AI also feeds on the labor of thousands of people. People who label millions of images, clean databases, moderate violent or sexual content, transcribe audio, translate texts, review errors. People rendered invisible by narratives of "total automation." People contracted through opaque platforms, without labor rights, often in Global South countries. Data workers, moderators, invisible curators are those who sustain the promise of functional AI. Without this human labor, artificial intelligence is not possible. So we ask ourselves again: Who sustains this development? In exchange for what? And for whom? Telling this story is an act of resistance. Because making visible what the hegemonic narrative silences allows us to imagine other possibilities. It allows us to question the current model, centered on the concentration of power, on the exploitation of bodies and resources, on technological acceleration without justice. It also invites us to build more just, more conscious, more situated technologies.‍

At the same time, we also imagine another artificial intelligence. While we denounce the invisible impacts of artificial intelligence development, we also propose something more: a different artificial intelligence. Because AI is not just a set of algorithms or "intelligent" machines. AI is a political phenomenon. It is constructed from human decisions, institutions, economic frameworks, and logics of power. It is shaped by who designs it, who finances it, from which languages, what data, and for what purposes. And for that very reason, it can be constructed in other ways.
From Diversa, we believe that another AI is possible.
A situated AI, which does not pretend to universalize its logic, but rather starts from local contexts and historically marginalized voices.
A feminist AI, which recognizes interdependence, which does not hide the human labor behind the technical process, and which places life, not economic profit, at the center.
A decolonial AI, which dismantles the hierarchies of knowledge and power that sustain dominant technology, and which builds from other territories, languages, and forms of knowledge.
A participatory AI, in which communities affected by its decisions have agency in its design, governance, and evaluation.‍

Our fanzine was part of the FemTech + Feminist Tech Exhibition, held on June 4, 2025, at the KTH Reactor Hall (Stockholm), a space for the intersection of art, science, and technology. The exhibition brought together interdisciplinary works that address intimate health, data, and care from feminist and decolonial perspectives. In this context, our fanzine was presented as a critical intervention on the materiality of artificial intelligence, making visible the bodies, resources, and territories that sustain its development. Being part of this exhibition allowed us to share these reflections with an international community of artists, researchers, and technologists committed to imagining more just technological futures. We hope to soon share this work in other languages and continue making these realities more visible.

‍Thank you for sustaining, reading, and making visible with Diversa.