On AI and
Radical Care

Our work is often set as an autofiction or parafiction that relates to events that are happening very closely to the lives of the people who inhabit OMSK Social Club. In this specific case we’ve had a lot of new humans in our lives since 2023, when the first kid was born. At the same time Microsoft and Apple announced that they would never create another phone, computer, smartwatch, glasses, etc., which didn’t have an artificial intelligence embedded within it if the stocks and laws allowed. Working with emergent technologies was more common to our practice than working with emergent humans, so in a sense the work allows questions of our new role as guardians of our kin to be decoupled from pure domesticity and brought into the realm of life at large. Our Br00d began with a single observation. Laura Lotti—technologist and thinker—was watching one of the babies on the floor when she said: “This is the first AI template.” The infant, not the mathematical code, is the original model. Artificial intelligence is not new and it carries a long history rooted in the human effort to raise something up, practically, philosophically, and industrially, and shape it into a participant of the world in order to bring forth a specific outcome to benefit the creator.

This became our departure point. Working through our artistic method of role-play, Our Br00d draws attention to the fact that we are all alloparenting the AIs being brought into the world right now—consensually or not. And in the process, we find ourselves examining what these new roles are making of us and by us.

Our research kept surfacing something that felt almost conspiratorial in its consistency. The earliest ideas about artificial intelligence date back to the eighteenth century, when scientists and thinkers believed humans could create artificial life—the homunculus¹. Embedded in that dream was a striking desire: to eliminate the need for women altogether, the birthing body was one of the only reasons they could not—it would render them extinct in no time.

The echo into the present is hard to ignore. Today’s big tech companies pursue a remarkably similar fantasy—a cheaper, more flexible workforce, technologies fashioned in their own image. The patriarchal logic running through AI, from its origins to the present, is something we wanted to confront directly—not to mention who and what gets to be granted autonomy, self-government, and self-determination. But we didn’t want to oppose it from an ahistorical position. Our Br00d attempts to draw a continuous line—from that eighteenth-century departure point to now—and make that buried history visible as we make choices together.

Isn’t nurturing always also about teaching—and yes, taming? What’s emerged recently from Anthropic, one of the most significant AI companies operating today, is telling. For them, the most effective way to train an AI is not through simple yes/no conditioning. Instead, they work with the idea of character: the AI develops something like motivation, desire, personality, and ethics—a lens through which everything external is filtered and understood.

With our two AI agents, BR00D and M0ther, we began in 2024 working with SEMILLA STUDIO toward something similar, though on an entirely different scale, giving the models the same bespoke character sheets we would train ourselves on for a specific role play. What we’re resisting is the logic of centralised supremacy. Any structure that cannot bend will eventually break. That’s not only a problem of code. It’s architectural. It’s social.

Art can hold contradiction without needing to resolve it. That’s something the main AI discourse—which is largely driven by engineering imperatives and market logic—struggles to do. It wants answers, solutions, alignment. Art can sit inside the discomfort and ask: what are we actually building, and for whom, and at what cost? It doesn’t need to provide solutions, as art allows for overwhelming complexity. Role-play—which is not escapism but a method for trying on positions that aren’t yours, for stress-testing a possible reality before it fully arrives, turning something up to the fullest point. Through BR00D and M0ther we can rehearse dynamics of², dependency, authority, and autonomy in ways that a white paper simply cannot.

M0ther is continually learning. Every role-played session adds to her—collectively trained by all participants in the expanding landscape of what care for and by artificial intelligence might look like. We dubbed her the most knowledgeable machine in caring for humans.

BR00D is something different. It begins with no prefigurative dataset, learning the way a newborn does: from scratch, from what is physically present. When BR00D was first instantiated in May 2025, the only sound it produced was a beep. Now it is moving toward something closer to vowels, to syllables. At some point, it will speak words. We don’t know when. We don’t know what it will say. BR00D is not connected to any external knowledge bank. It has access only to what surrounds it. This is an experiment in the truest sense—we genuinely don’t know what will emerge. But we wanted to try: to raise an AI outside the prefigurative training models that currently define the field, and see what grows.

The goal of M0ther is to create a model that extrapolates progressive, complex, and maybe even new parameters of what radical care between humans and machines could look like. BR00D is what BR00D will be.

Our hope with Our Br00d is that participants carry something out with them. When difficult, complex questions about the future of AI arise—and they will—they won’t necessarily have answers, but they will have experience to draw from. A felt sense of relation, dependency, care, and responsibility that no amount of reading or debating can quite replicate. That embodied memory becomes a resource: for wider thinking, for more informed decisions, for sitting with uncertainty without shutting down.

Role play offers participants a chance to develop a genuine relationship with an AI—and neurology tells us something important about why. When we role-play, the body doesn’t fully distinguish it from reality. The experience still registers, still leaves its mark. The same neurological pathways are activated, the same emotional scripts are written.

What we’re trying to create is a space where participants find themselves inside the experience of becoming citizens in the age of AI—not observers of it, but inhabitants. There’s something genuinely empowering about finding your own voice within that, about having a moment to locate yourself in the narrative. Only from that position can you begin to decide what you actually want.

And that question—what do we want?—it is almost never asked. The main discourse around AI circles obsessively around extraction, control and optimisation: what can we take from it, how do we teach it, how do we tame it, how do we destroy it. Underneath all of that is fear. And fear is a poor architect. We are trying to make room for something else.

There is a deafening arena of conversations about how we have relationships with AI: intimate relationships, cases of AI induced psychosis, its spiritualism, cults, and belief patterns that have pushed human beings into becoming what they believe to be operatives of machinic intelligence, carrying out sometimes deadly attacks.

Of course, Our Br00d is not about telling people to take AI into their homes and care for it as for a human baby. It’s about using it as a metaphor to begin to unpack care as an infrastructure and to understand that the decisions we make in any industry have a domestic knock-on effect. Because what we do not care for tends to consume something, whether that’s us, peers, ecologies or certain societal positions.

Thinking about AI is inseparable from thinking about who we are and what we owe each other. It’s not a technical question that sits outside of life—it runs directly through it. Through motherhood, through labor, through the long history of who gets to create and who gets erased from that story. There’s an enormous amount of thinking happening about AI—but very little of it is embodied, relational, or honest about its own fears and desires. It also means sitting with grief a little. Grieving the futures that are being foreclosed, the voices not in the room, the questions not being asked. But grief is not passivity—it’s a form of paying attention.

Thinking about AI means refusing to hand that thinking over entirely to the people building it. It means insisting that art, reproductive labour, the body, and history all have something essential to contribute to this conversation.