How to Creatively Use AI in Images — Blind Journal

Hana Katoba approaches image-making like a traveler returning to a well-recognized panorama solely to search out it reworked. Trained as a photographer and formed by years of observing mild, gesture, and environment, she entered the realm of AI with out the urgency of novelty. Instead, she treats it as an extension of her visible apply: one other room in the home she has been constructing since she first started taking footage. The digicam is now not a bodily object however an area the place instinct and reminiscence information the composition.

For Katoba, synthetic intelligence isn’t a shortcut however a medium of exploration. She sketches ideas as she as soon as would with a lens, transferring by prompts the way in which she as soon as moved by a metropolis or arrange—slowly, letting uncertainty lead. The ensuing photographs carry the identical chromatic delicacy that marked her work: dreamlike portraits, spectral architectures, fragmented landscapes that echo the surreal with out abandoning coherence. The fantasy in her compositions isn’t escapism; it’s a strategy to contact the emotional residue of actuality.

In her business collaborations, Katoba adapts this language to the wants of manufacturers and editorials, with out relinquishing authorship. She sees the core problem not in utilizing AI, however in retaining one’s voice inside it. The expertise can multiply potentialities, she says, however solely the artist can set course.

In the next interview with Blind, Hana Katoba explains how she navigates that area—between images and invention, self-discipline and play—and the way AI has develop into much less a software than a companion shaping her visible creativeness.

How did you first start integrating AI into your photographic apply? 

I began integrating AI very organically, virtually out of curiosity. I had been working in images for a few years, and after I found the probabilities of generative fashions, I felt I might broaden my visible language past what the digicam and my assets allowed. It wasn’t a lot an abrupt change because it was a pure transition: a strategy to translate concepts that beforehand solely existed as sensations or psychological sketches. 

Was there a specific challenge or second if you realized AI might play a central position in your visible language? 

There wasn’t one particular challenge that marked the second; it was extra of a gradual revelation. While working with AI, I found I might rework very summary concepts into photographs with out the technical and budgetary limitations of conventional media. That new, borderless area allowed my creativeness to broaden with whole freedom. What actually made me perceive its potential was realizing that the method woke up a manner of making that felt very near play. I deeply imagine that, as people, studying by play is innate; it’s how we develop, discover, and uncover the world for the primary time. With AI, I recovered that primal impulse: to experiment with out worry, to comply with my instinct, to be shocked. That mixture of freedom and play—that risk of constructing from essentially the most important curiosity—was what made me perceive that AI might develop into a central a part of my visible language.

How does your use of AI differ between your private tasks and your business assignments? 

In my private tasks, I permit myself a freer, extra emotional exploration, the place AI acts as an area for introspection. In the business realm, my use of AI takes on a extra polished and strategic character. It’s not a lot about letting go, however about translating ideas into photographs that dialogue with the shopper’s imaginative and prescient with out shedding my visible identification. That course of might be difficult as a result of there’s a danger that your individual voice may get diluted amongst so many exterior tips. For me, the fascinating half is managing to offer sensitivity, authenticity, and a private perspective even inside a tighter framework.

In business work, what expectations or boundaries do shoppers usually have relating to AI-generated or AI-enhanced imagery? 

Clients normally search for innovation, however with clear limits: aesthetic coherence, respect for model identification, and technical viability. Many have an interest within the narrative or conceptual potential of AI, however they want the consequence to be managed, reproducible, and aligned with the challenge’s goal.

When you begin a picture involving AI, what’s your entry level — {a photograph}, a immediate, a sketch, an idea? 

Normally, I begin with an idea. Sometimes I develop it into a fast sketch or a written description that serves as a place to begin. Little by little, I unravel all of the traits that can form the picture: how the composition will probably be organized, what sort of mild matches finest, and what colour palette can maintain the environment I’m in search of. That start line helps me preserve coherence after I later incorporate AI into the method. 

But what actually fascinates me is the exploration of latent area. It’s a technical time period, sure, however I’d clarify it like this: think about with the ability to journey from an preliminary picture or concept towards some other visible vacation spot, and through that journey, surprising findings seem, identical to if you journey with out a mounted plan. For me, producing photographs with AI turns into that journey: I can change course, experiment, uncover, and take management at any second. Each picture turns into a unique station inside an infinite artistic route.

Could you stroll us step-by-step by your typical workflow for creating an AI-assisted picture, from concept to last consequence? 

● Idea / Concept: Everything begins with a transparent intention: an emotion, an environment, or an concept I need to handle. 

● Brief Description: I normally make a small sketch or write a number of traces defining the course. It doesn’t must be elaborate, however sufficient to set the “mental framing,” as if I have been planning a photograph shoot. 

● First Iterations with AI: From there, I construct the prompts, holding elementary ideas very current: lighting kind, fashion, colour, angle. I check totally different phrase combos and observe how the pictures reply. 

● Journey by Latent Space: Here begins essentially the most stimulating a part of the method. I like to think about latent area as an infinite map the place all doable photographs are related to one another. Exploring that map is like touring with out a mounted vacation spot: I begin from an preliminary concept, and whereas testing iterations, I uncover surprising paths. I can deviate, return, or take a totally new course, at all times guided by the unique emotion. Each generated picture is a unique cease on that artistic journey. 

● Image Selection: Of all of the variants, I select the one that isn’t solely “correct” on a technical degree however the one which finest captures the preliminary emotion or concept. It’s a course of

similar to enhancing a images sequence: it’s about discovering the picture that sustains the message. 

● Editing / Post-production: Afterward, I work on the picture in post-production as I’d with {a photograph}: tone, colour, distinction, texture, depth notion. This half is vital to unifying the piece with the remainder of my work so it isn’t perceived as one thing remoted. 

How do you construct prompts in a manner that retains the outcomes private and aligned together with your aesthetic slightly than generic? 

What retains my outcomes private isn’t the immediate, however my perspective (“mirada”) behind the entire course of and the journey. Honestly, the immediate isn’t an important half in any respect; it’s simply an preliminary level, a tough sketch. As an entire, that mixture—my manner of seeing, my visible language, and my photographs as a information—prevents the consequence from being generic and ensures each bit maintains coherence with my work. 

Which AI fashions, software program, or platforms are central to your course of immediately, and why? 

My major software is ComfyUI. It might be described because the “operating system” for AI-generated imagery: an atmosphere the place I can join totally different fashions and, on the identical time, management intimately each parameter concerned in setting up a picture or video. In my private work, I generally permit for extra experimentation and probability, however in business tasks, that degree of management turns into important: with the ability to alter particular particulars, reply to concrete adjustments, and assure a precise result’s a part of the skilled course of.

How do you handle decision, texture, and effective element — frequent limitations of AI-generated photographs? 

I generate the bottom picture, and after I discover the model that works, I upscale it to achieve high quality, sharpness, and realism. From there, I intervene manually in post-production: I alter texture, mild, and element. That last work is what ensures the picture has visible coherence and an natural really feel, far faraway from the unreal.

What are the primary technical pitfalls that newcomers ought to keep away from when integrating AI right into a photographic workflow? 

One of the commonest errors is trusting utterly in what the AI generates and setting apart one’s personal judgment. It additionally occurs that, with so many choices out there, it’s simple to lose aesthetic readability: both by including too many results or by following descriptions too actually. It’s not that these components are an issue in themselves, however they’ll divert consideration from what issues most: the intention of the picture. The key lies in sustaining a transparent course, working with persistence, and remembering that AI is a software inside a broader artistic course of.

What steadiness do you goal for between AI-generated components and conventional photographic or post-production work? 

AI intervenes in my course of, however it doesn’t outline the picture. What actually sustains each bit is my gaze as a photographer: the way in which I perceive mild, colour, area, and the themes I discover. I don’t delegate these selections; they arrive from years of coaching and an eye fixed educated to construct photographs with intention. AI helps me generate preliminary materials or open up new visible potentialities, however I’m the one who marks the course from the beginning and who, later, provides cohesion to the work in post-production. In brief, AI expands the method, however the visible identification comes from me and is current at each stage, not simply within the last end.

Do folks ever misunderstand what “AI photography” means within the context of your work? 

Yes, it occurs incessantly. When they see a generated picture that preserves sure photographic codes—the sunshine, the composition, the depth—they have a tendency to interpret it as an actual scene, as if it had existed in entrance of a digicam. But in my case, there is no such thing as a captured prompt: there’s an imagined scene. AI permits for constructing photographs that lean on the visible language of images however don’t have a tangible origin. It’s a manner of giving visible kind to that which usually stays within the thoughts: concepts, feelings, fuzzy recollections… and even one thing much like capturing a dream proper earlier than it fades away.

Your photographs have a really distinctive environment. How do you preserve a coherent fashion when working with instruments that may be generally unpredictable? 

The coherence of my photographs comes from my background and my manner of understanding images. Light, colour, composition, and environment aren’t random components to me: they’re selections I’ve been refining for years. Although AI might be unpredictable, my standards acts as a relentless filter. I assessment, alter, and choose till the picture breathes my language.

How do you protect the emotional and sensory dimension of images when utilizing algorithmic instruments? 

I attempt to make emotion the place to begin and never the consequence. AI helps me construct atmospheres, however I’m the one who decides the place to put the silence, the strain, the humor. I consider every picture as a small area the place the viewer can cease and really feel.

Could you describe a private challenge the place AI considerably reworked your preliminary concept, and clarify the way you constructed the ultimate picture? 

It’s not one thing that occurred in only one particular challenge, however in all these the place I permit myself to experiment. Sometimes I begin with a really primary information and journey by the latent area down surprising paths. That openness normally leads me to outcomes which can be far more artistic than I imagined. It occurs particularly in photographs with a humorous or surreal tone, the place there’s room to play and let extra “out of the box” visible options seem.

If you have been educating somebody new to AI, what’s one concrete method you’d suggest they begin with? 

I’d invite anybody beginning out to create a small “style moodboard”: compile 5 of their very own photographs that signify their visible style and use them as a recurring reference. This manner, you be taught from the start to construct aesthetic coherence and never rely solely on what the mannequin proposes. But, above any method, I’d invite them to take care of an angle of play: experiment and let your self be shocked. Creativity with AI is discovered by enjoying.

How do you see AI reshaping the position of photographers within the coming years? 

I believe it should broaden the photographer’s position, not exchange it. It will flip us into designers of worlds, not simply photographs. The digicam will now not be the one medium to seize, however considered one of many to construct. The important factor will proceed to be the gaze. 

Are there new instruments or applied sciences you’re excited to experiment with subsequent? 

I’m particularly excited concerning the arrival of real-time world-generation instruments. We are already seeing advances in video fashions, getting nearer to that concept, however we nonetheless lack the required immediacy. I think about a really close to future the place we will transfer from one picture to a different—from one visible universe to a different—fluidly, modifying any component immediately. When that occurs, the artistic expertise will probably be virtually like inhabiting a dream that we will design and management within the first individual. An area the place not solely photographs are generated, however complete realities to work together with. And when that expertise arrives… I’ll most likely get misplaced in there, so should you search for me, you understand the place to search out me.

Looking forward, what place do you think about AI can have in your creative apply in 5 or ten years? 

I think about a future the place “traveling” by latent area isn’t only a metaphor, however a literal expertise: with the ability to transfer from one picture to a different, from one environment to a different, inside a visible universe in fixed transformation, the place any component might be modified immediately. We gained’t simply generate photographs anymore: we’ll inhabit them. That leap will flip creation into one thing immersive, virtually like daydreaming and having whole management over the dream. Current instruments are already pointing that manner, however when that expertise matures, it should utterly change my course of… and doubtless my relationship with the picture too. The thrilling factor is that it’s arduous to think about how far we’ll go; the tempo of evolution is dizzying. But if one factor is evident to me, it’s that I’ll be there, exploring it, enjoying… and fortunately getting misplaced in these worlds.

More info on Hana Katoba on her website or Instagram. Hana Katoba is represented by FMA Le Bureau.