Generated Identities
Role of artificial intelligence, fashion, and the construction of identity.
Aadhya Khare · 31st March 2026
Decoding the Zeitgeist
I've been thinking about how much of our lives we are beginning to outsource to artificial intelligence. We ask it what to wear, how to look, what suits us, what we should say, and sometimes even how we should feel. Slowly, we are inviting an external intelligence into decisions that once came from instinct, experience, and personal preference. Fashion makes this particularly interesting because clothing is one of the ways we construct our identities. What we wear says something about who we are, who we want to be, and how we want the world to perceive us. But what happens when the choices through which we express ourselves begin coming from somewhere else? At what point does self-expression become a collection of generated suggestions that we have simply learnt to accept? I was interested in the cultural shift beneath the convenience of AI: our growing willingness to let external systems participate in constructing who we are. The concern wasn't that technology could influence our choices, but that we might gradually stop questioning where those choices come from. That was the question I wanted to sit with.
Psychological & Archetypal Decoding
What interested me psychologically was the kind of person who might find comfort in having their identity interpreted, organised, and validated by something outside themselves. I began thinking about the perfectionist archetype, particularly its less obvious side. Perfectionism can look like confidence, discipline, and an uncompromising eye for detail. But sometimes, underneath the desire to get everything right, there is a deep discomfort with uncertainty. The fear of making the wrong choice can become so consuming that someone begins looking outside themselves for reassurance. And what better system to turn to than one that can produce an answer for almost anything? When external validation becomes the measure of whether a choice is right, personal preference can take a back seat. The individual may become exceptionally good at presenting a desirable version of themselves while becoming increasingly uncertain about what they actually desire. These archetypes aren't diagnoses, nor are they mutually exclusive identities. They are psychological patterns through which I wanted to examine the relationship between insecurity, control, validation, and self-expression. What makes this relationship particularly interesting is that the person may not feel as though they are losing their identity. They might feel as though they are improving it. Every recommendation makes the next decision easier. Every perfectly assembled outfit offers reassurance. Every generated image provides another reference for who they could become. But when does the pursuit of a better version of yourself become a way of avoiding the uncertainty of being yourself? This became the psychological tension at the heart of Generated Identities: the possibility that our need for certainty could make us increasingly dependent on systems that tell us who we should be.
Core Concept Development
I called the project Generated Identities because I wanted to explore more than the images artificial intelligence can produce. I was interested in the identities we might begin producing through our dependence on it. An identity can be assembled from references, trends, recommendations, and visual codes. It can be refined until every element looks right, made coherent, desirable, and perfectly presentable. But does looking like a complete person mean feeling like one? That distinction became central to the project. I wanted to examine the distance between an identity that is lived and one that is constructed to appear convincing. Fashion is deeply personal, yet the systems influencing our choices increasingly operate through patterns, probabilities, and standardised outputs. A machine can suggest what is desirable, suitable, or visually coherent, but it cannot fully account for the contradictions that make an individual human. I therefore approached the work as an observational art piece rather than a literal narrative about someone interacting with technology. I wasn't interested in showing a screen, a prompt, or a machine making decisions. I wanted to explore the human consequences of that relationship through the very thing AI can produce so convincingly: the image.
Visual World Development
I deliberately approached the imagery through polish, visual control, and carefully constructed presentation. I wanted the images to look complete. The colours, textures, silhouettes, styling, casting, and overall visual treatment needed to create a sense of refinement, as though every element had been considered and placed exactly where it belonged. The styling was particularly important. The garments worked together, the combinations made sense, and the silhouettes felt cohesive. Everything was visually resolved. But that was precisely the point I wanted to question. When we tell an AI system about our style preferences, the trends we gravitate towards, and the references that influence us, it can interpret those inputs and assemble an outfit that fits the brief. It can match colours, balance proportions, combine references, and produce something that looks as though every decision has been made correctly. But an outfit that works is not necessarily an outfit that expresses you. There is a difference between wearing something because it reflects an instinct, a memory, a contradiction, or a feeling, and wearing something because all its elements happen to work together. One can emerge from a deeply personal relationship with clothing; the other can be assembled through a set of successful visual calculations.
The Styling
I wanted the styling to sit within that tension. The outfits weren't supposed to look badly put together. Their polish was essential because it allowed me to question whether visual coherence could be mistaken for personal expression. The same thinking extended to the colour palette, textures, and silhouettes, creating a world that felt deliberate and controlled. The casting and presentation positioned the individual as someone whose identity could be read through an image, while the lighting and composition reinforced the sense of a constructed persona. A polished image can hold our attention through its formal qualities, even when it offers very little emotional intimacy. Everything can appear resolved on the surface while something remains absent underneath. I wanted that absence to matter. The visual world was designed around a contradiction: the more complete the image appeared, the more I wanted the viewer to question what it was actually communicating or have we stopped asking that altogether?
Experience Direction
I wanted the audience to encounter the work before being told what to think about it. Approaching Generated Identities as an observational art piece allowed the imagery to exist without a conventional storyline explaining every decision. Rather than guiding the viewer towards a predetermined emotional response, I wanted to leave room for them to examine their own relationship with the images. Movement, pacing, and the treatment of the subject became important to this experience. The body communicates through posture, gesture, repetition, and stillness. These details influence whether a person feels spontaneous, vulnerable, controlled, or carefully constructed. In a project concerned with generated identity, movement becomes another way to question the difference between something that looks convincing and something that feels lived. A composed image can suggest a complete persona, but the presence of a body also reminds us that identity exists beyond appearance. I didn't want the work to become a demonstration of what AI can do. I wanted it to become an encounter with what its increasing presence might mean for us. And I wanted the viewer to arrive at the question themselves: am I responding to this image because it makes me feel something, or because I recognise that it looks good?
Creative Direction
The final film extended this investigation into time, rhythm, movement, and visual continuity. While a still image captures a constructed identity at a particular moment, moving imagery allows us to observe how that identity is presented and sustained. I was interested in what happens when the visual experience remains polished but emotionally distant. The film brings together fashion, constructed identity, and AI-generated imagery to examine the difference between producing something that looks complete and creating something that makes us feel something. The use of AI-generated imagery also made authorship part of the question. Artificial intelligence can generate a convincing representation of a person without possessing that person's lived experience. It can reproduce the appearance of emotion without having experienced the emotion itself. This distinction matters to me because art has always been about the capacity to make us feel, question, remember, recognise, or confront something within ourselves. It can unsettle us, make us uncomfortable, or remind us of something we cannot quite articulate. A visually perfect image does not automatically do any of these things.
And this brings me to the question I really want to ask anyone who encounters this work. When you look at these pictures, do you feel anything? Maybe you don't. Maybe you admire the images. Maybe you find them beautiful, polished, or visually interesting, but feel no particular attachment to them. That possibility was central to the thought behind Generated Identities. I wanted to question what happens when the production of an image becomes detached from the emotional experience that makes art meaningful to us. Artificial intelligence can generate images that appear finished, refined, and convincing. But visual completion and emotional connection are two very different things. Generated Identities was my way of sitting with that contradiction. The work explores the distance between looking and feeling, between constructing an identity and experiencing one, between an image that appears human and the human experience that may be missing from it. Perhaps the most unsettling possibility is that we could become so accustomed to admiring generated images, accepting generated suggestions, and constructing generated versions of ourselves that we stop questioning what we are actually connecting with.And perhaps that is where the title extends beyond the images themselves. If an identity can be generated, an image can be generated, and even the appearance of emotion can be generated, what remains that we can genuinely call our own?
Psychological & Archetypal Decoding
What interested me psychologically was the kind of person who might find comfort in having their identity interpreted, organised, and validated by something outside themselves. I began thinking about the perfectionist archetype, particularly its less obvious side. Perfectionism can look like confidence, discipline, and an uncompromising eye for detail. But sometimes, underneath the desire to get everything right, there is a deep discomfort with uncertainty. The fear of making the wrong choice can become so consuming that someone begins looking outside themselves for reassurance. And what better system to turn to than one that can produce an answer for almost anything? When external validation becomes the measure of whether a choice is right, personal preference can take a back seat. The individual may become exceptionally good at presenting a desirable version of themselves while becoming increasingly uncertain about what they actually desire. These archetypes aren't diagnoses, nor are they mutually exclusive identities. They are psychological patterns through which I wanted to examine the relationship between insecurity, control, validation, and self-expression. What makes this relationship particularly interesting is that the person may not feel as though they are losing their identity. They might feel as though they are improving it. Every recommendation makes the next decision easier. Every perfectly assembled outfit offers reassurance. Every generated image provides another reference for who they could become. But when does the pursuit of a better version of yourself become a way of avoiding the uncertainty of being yourself? This became the psychological tension at the heart of Generated Identities: the possibility that our need for certainty could make us increasingly dependent on systems that tell us who we should be.
Visual World Development
I deliberately approached the imagery through polish, visual control, and carefully constructed presentation. I wanted the images to look complete. The colours, textures, silhouettes, styling, casting, and overall visual treatment needed to create a sense of refinement, as though every element had been considered and placed exactly where it belonged. The styling was particularly important. The garments worked together, the combinations made sense, and the silhouettes felt cohesive. Everything was visually resolved. But that was precisely the point I wanted to question. When we tell an AI system about our style preferences, the trends we gravitate towards, and the references that influence us, it can interpret those inputs and assemble an outfit that fits the brief. It can match colours, balance proportions, combine references, and produce something that looks as though every decision has been made correctly. But an outfit that works is not necessarily an outfit that expresses you. There is a difference between wearing something because it reflects an instinct, a memory, a contradiction, or a feeling, and wearing something because all its elements happen to work together. One can emerge from a deeply personal relationship with clothing; the other can be assembled through a set of successful visual calculations.
Experience Direction
I wanted the audience to encounter the work before being told what to think about it. Approaching Generated Identities as an observational art piece allowed the imagery to exist without a conventional storyline explaining every decision. Rather than guiding the viewer towards a predetermined emotional response, I wanted to leave room for them to examine their own relationship with the images. Movement, pacing, and the treatment of the subject became important to this experience. The body communicates through posture, gesture, repetition, and stillness. These details influence whether a person feels spontaneous, vulnerable, controlled, or carefully constructed. In a project concerned with generated identity, movement becomes another way to question the difference between something that looks convincing and something that feels lived. A composed image can suggest a complete persona, but the presence of a body also reminds us that identity exists beyond appearance. I didn't want the work to become a demonstration of what AI can do. I wanted it to become an encounter with what its increasing presence might mean for us. And I wanted the viewer to arrive at the question themselves: am I responding to this image because it makes me feel something, or because I recognise that it looks good?
And this brings me to the question I really want to ask anyone who encounters this work. When you look at these pictures, do you feel anything? Maybe you don't. Maybe you admire the images. Maybe you find them beautiful, polished, or visually interesting, but feel no particular attachment to them. That possibility was central to the thought behind Generated Identities. I wanted to question what happens when the production of an image becomes detached from the emotional experience that makes art meaningful to us. Artificial intelligence can generate images that appear finished, refined, and convincing. But visual completion and emotional connection are two very different things. Generated Identities was my way of sitting with that contradiction. The work explores the distance between looking and feeling, between constructing an identity and experiencing one, between an image that appears human and the human experience that may be missing from it. Perhaps the most unsettling possibility is that we could become so accustomed to admiring generated images, accepting generated suggestions, and constructing generated versions of ourselves that we stop questioning what we are actually connecting with.And perhaps that is where the title extends beyond the images themselves. If an identity can be generated, an image can be generated, and even the appearance of emotion can be generated, what remains that we can genuinely call our own?
Decoding the Zeitgeist
I've been thinking about how much of our lives we are beginning to outsource to artificial intelligence. We ask it what to wear, how to look, what suits us, what we should say, and sometimes even how we should feel. Slowly, we are inviting an external intelligence into decisions that once came from instinct, experience, and personal preference. Fashion makes this particularly interesting because clothing is one of the ways we construct our identities. What we wear says something about who we are, who we want to be, and how we want the world to perceive us. But what happens when the choices through which we express ourselves begin coming from somewhere else? At what point does self-expression become a collection of generated suggestions that we have simply learnt to accept? I was interested in the cultural shift beneath the convenience of AI: our growing willingness to let external systems participate in constructing who we are. The concern wasn't that technology could influence our choices, but that we might gradually stop questioning where those choices come from. That was the question I wanted to sit with.
Core Concept Development
I called the project Generated Identities because I wanted to explore more than the images artificial intelligence can produce. I was interested in the identities we might begin producing through our dependence on it. An identity can be assembled from references, trends, recommendations, and visual codes. It can be refined until every element looks right, made coherent, desirable, and perfectly presentable. But does looking like a complete person mean feeling like one? That distinction became central to the project. I wanted to examine the distance between an identity that is lived and one that is constructed to appear convincing. Fashion is deeply personal, yet the systems influencing our choices increasingly operate through patterns, probabilities, and standardised outputs. A machine can suggest what is desirable, suitable, or visually coherent, but it cannot fully account for the contradictions that make an individual human. I therefore approached the work as an observational art piece rather than a literal narrative about someone interacting with technology. I wasn't interested in showing a screen, a prompt, or a machine making decisions. I wanted to explore the human consequences of that relationship through the very thing AI can produce so convincingly: the image.
The Styling
I wanted the styling to sit within that tension. The outfits weren't supposed to look badly put together. Their polish was essential because it allowed me to question whether visual coherence could be mistaken for personal expression. The same thinking extended to the colour palette, textures, and silhouettes, creating a world that felt deliberate and controlled. The casting and presentation positioned the individual as someone whose identity could be read through an image, while the lighting and composition reinforced the sense of a constructed persona. A polished image can hold our attention through its formal qualities, even when it offers very little emotional intimacy. Everything can appear resolved on the surface while something remains absent underneath. I wanted that absence to matter. The visual world was designed around a contradiction: the more complete the image appeared, the more I wanted the viewer to question what it was actually communicating or have we stopped asking that altogether?
Creative Direction
The final film extended this investigation into time, rhythm, movement, and visual continuity. While a still image captures a constructed identity at a particular moment, moving imagery allows us to observe how that identity is presented and sustained. I was interested in what happens when the visual experience remains polished but emotionally distant. The film brings together fashion, constructed identity, and AI-generated imagery to examine the difference between producing something that looks complete and creating something that makes us feel something. The use of AI-generated imagery also made authorship part of the question. Artificial intelligence can generate a convincing representation of a person without possessing that person's lived experience. It can reproduce the appearance of emotion without having experienced the emotion itself. This distinction matters to me because art has always been about the capacity to make us feel, question, remember, recognise, or confront something within ourselves. It can unsettle us, make us uncomfortable, or remind us of something we cannot quite articulate. A visually perfect image does not automatically do any of these things.