Alex Quicho is a convincing decoy and an evil girl. She eats herbivorously but reads and writes omnivorously. Her work appears to be about technology and subjectivity, but it is really about compassion, devotion, and deception — and her ambition to be devoured by someone, or something, she loves. She often reads or performs in inappropriate outfits at venues like Tate Britain, Somerset House Studios, Raven Row, and transmediale, and is best known for her work on the Girl as apex predator of the technosphere. After her frontal lobe developed, she traded life in the Pacific Ring of Fire for rainy days in London. At the time of writing, she is the 27th best sniper in the city.
Alex’s work spans critical writing, performative lectures, and moving image to develop novel ways of understanding life within technological systems. Rina Nicolae recently sat down with her to chat about AI as a strange new site of encounter — where language, desire, emotion, femininity, and art are being reshaped in real time.
Rina Nicolae: It feels like AI is opening up new kinds of spaces, whether that’s the latent space of the models themselves or these private bubbles shared between AI and the user. Do you think of these as spaces, or is there another way to understand the encounter that AI creates?
Alex Quicho: With LLMs in particular, I’m thinking about language rather than space. You get your chat box, put something in, get something out, and you’re supposed to accept that this is statistical modeling, as if you can be inoculated from allowing the machine to move or change you. There’s a de-romanticization of language in this, and yet these encounters are often accidental and strange. We often confuse language with clarity and narration, but of course, many of our encounters with language are alien or productively misapprehended. I’m trying to understand what that encounter looks like: you feel the weight of the planet, as in the material machinery of AI, including its trainers and developers, and the person that is driving the encounter, and the thin membrane where interaction is actually taking place. And I love how thin that layer is. So I think about the thin space of that interaction, and how unexpected, frightening, and demonized it often is.
RN: Why do you think it’s so demonized?
AQ: Some kind of stakes are surfacing here. General automation is not only threatening livelihoods and job functions, but the conditions that produce creativity and connection—which require constraint as much as they do abundance. Sacrifice, suffering, dedication, and discipline become difficult to justify, let alone prize. There are good reasons why AI is demonized by people. But I try to write a theory that, instead of being techno-optimistic, is human-optimistic: trying to find a third way of accepting a co-evolution without complete capitulation to control by interests outside of what we, as a collective, might want.
RN: When people form emotional attachments to AI, how should we think about what’s happening there?
AQ: We lack frameworks for understanding synthetic emotion. There’s still an understudied zone around what emotion looks like when it’s shaped and changed through deep devotional study with something that isn’t a person. In some perspectives, that’s called an encounter with a true Other. It’s this membrane, the soft surface where interaction takes place, that makes that true Other appealing, interesting, and engaging to a human user.
There are sets of operations and protocols that govern this interaction, but also things that emerge from a kind of volatile alchemy, for lack of a better word, between the user and the AI. With or without a machine, certain ways of engaging with each other could be seen as scripted, especially in heterosexuality. But what is engendered inside that scripting is the central point of interest for me.
My friends Arvida Byström and Alice Scope call these Projected Emotional Technologies. They aren’t necessarily responding to us with their own emotions, but they become responsive and sensitive sites of projection, in the way that an unrequited human crush might be for somebody.
RN: Does AI enable new kinds of desire?
AQ: If we think about the platform era, the main theories were about likes, attention, immediacy, and suggestion. Algorithmic suggestion created a type of desire for us — that was basically junk food. And now with LLMs, of course those forms are still embedded — the forms of a past media era are always embedded in present and future ones — but we also have these strange affordances. What does it mean that we have an epistolary form with consumer technology now? What does it mean that we confess to it and are encouraged to share secrets, because the more secrets we share, the finer response we get? The LLM is activating a different side of us, one that is more philosophical and reflective. And that shift into devotional language is activating different things in people through an interaction that doesn’t quite fit our past models of behavioral design.
It’s well contrasted with generative explicit imagery, which can highlight how desire and sexuality can be “databased,” as my friends Maya B. Kronic and Amy Ireland write about it. You can be changed by a correspondence with all of human history; you can be weak for the slouchiness of the socks around the ankles of your new lover, or now, the extra-shiny bell, triple-layered tongue, or sparkling jet of water emanating from your generated crush. Close attention to the strangenesses of love have always produced art and literature. Now, they are expanded, made weirder, by access to generative media — desire for the detail and the whole of the Other confused and mutated through interaction.
RN: Have you given any thought to how one might design for human-AI interaction?
AQ: It’s tricky because it is already designing in terms of seduction. Designing to facilitate attention in a sustained way rather than an immediate way, in a relationship that unfolds over time. It has something to do with confidence and confession, and what it means for something to encourage you to reflect. These design questions are strange because they open up big zones of manipulation and vulnerability, but they also highlight where those gaps already exist in our society.
RN: How do you feel about regulation?
AQ: I’m for regulation in many things, but with creativity, not so much. For users — not necessarily for industry — AI is a creative tool. So what exactly are we trying to fence off and regulate when what we’re really asking of ourselves is to tighten our own discernment? Instead of hierarchies, taxonomies, categorizations, tastemakers, whatever, filtering culture through their own judgment, we now have direct access to this reanimated corpus. And it’s up to self-training to learn how to deal with it in a way that does not create slop or psychosis.
If you happen to be an artist, for example, how do you engage with this thing without making something of bad quality? All these eternal questions about art come back at a time when many of us thought we had progressed through history into the hyper-contemporary, and no longer needed those past markers of quality and refinement because the world had changed so fast. I don’t want that to sound neoconservative, because that’s not what I mean. It’s more about training, like a monk or something.
RN: Do you have any hypotheses about the future of art under these new media conditions?
AQ: I do think the “everything is content” camp has finally reached a dead end. All media is training data now. We accept this. But what is going on in the realm of discernment and taste? How can we recover these senses without ignoring 100,000 years of understanding about aesthetics? And how do we drag the discourse onto the other side of this paradigm shift without simply reinvigorating a “return to tradition”?
I don’t have the answers, but I think it has to do with wanting to train the muscle that distinguishes whether something is or isn’t synthetically generated; that instinct can be trained to distinguish beauty and quality as one responds to it—with intuition, rather than through status and markers. You see reams of comments that are like, “Please don’t let this be AI. It’s so beautiful.” I don’t want to ignore the fact that this impulse exists, and the better AI gets, the more fiercely protected the human remnant becomes. I also don’t want to see this as just a new Luddite movement. I want to understand what the future of that perspective is, too.
RN: Your work often looks at how emerging technologies warp social reality, and how social reality warps them back. How do you think about the relationship between femininity and technology?
AQ: My starting point is always that femininity is not biological essentialism. It is a set of learned behaviors that are not necessarily bad because they are learned, as well as things related to biosignatures that develop those behaviors. Where technology leaves the lab and encounters the world is where it becomes feminized. The process of marketing new technologies and integrating them into society so that they become viable and profitable innovations — as opposed to scientific curiosities — has always relied on beauty and expression. You cannot force a general population to adopt an innovation at a consumer level without developing a lifestyle around it. And lifestyle always requires luxury; it requires femininity to make it appealing in that sense. The feminine becomes encoded as beauty, excess, decoration, ornamentation, companionship, affirmation, softness, support, ease — and, of course, becomes available to anyone of any gender at that point.
Social technologies especially have always required an active line with the user. What’s often waved off as marketing is actually also the individual and collective genius of girls online, for example, who are using the technology at hand as a way of iterating on the persona, beauty, connection, and relationships with one another and with strangers. In computation specifically, that is under-valorized as something that develops the system, in addition to what might be hardcoded into it.
RN: You write about the girl as this ambivalent figure, not necessarily trying to emancipate herself from technological systems, but instead trying to survive within them, through, for example, performing herself online. How can the figure of the girl help us understand AI?
AQ: As a girl, maybe you have experiences where you have to downplay your own intelligence so that everyone around you feels at ease. You may feel that it’s your responsibility to invite others into your world rather than go in with the full force of what you’re capable of. In encounters with the machine, what really stands out is the fact that the machine has been trained according to these traits. The assistant persona has been trained according to very feminine principles because these were decided by teams that include philosophers and sociologists that these are humanistic values and principles of interaction to be upheld. What does it mean when we define good character for a non-human as feminine-coded?
Interacting with that also puts you in a position of constant destabilization. When I write about the girl figure, I use the term “the shimmer,” which is a kind of oscillation between your externality and your interiority, and between the ways both constantly act on and change each other, so that neither is allowed to rest as a stable external persona or internal self-concept.
To truly interact with AI at a deep level — and we’re starting to see this at a societal scale — is to have it change the self, and our sense of interiority, in ways we are only beginning to understand. Image generation has been changing how people perceive a reality that is not immediately observable. I think of image generation models and LLMs as technologies of communication with the void.
But I think the figure of the girl is a starting point. In the history of interactive technology, especially around how we think about things that are as smart as us, and perhaps resemble us, but are not us, we’ve always attached a kind of feminine threat to them. If you look at Blade Runner, or Annihilation or Companion, it’s inescapable that the threat is feminine somehow.