YTH | Art Direction and Emotional Narrative
AI tools can swiftly generate imaginative visuals that art directors incorporate into moodboards and concept art. Instead of spending days hunting for the perfect photograph or sketching concepts from scratch, art directors can now prompt an AI to produce a palette of options in minutes. This fundamentally changes early visual research: AI systems act as on-demand “imagination assistants,” outputting rapid iterations that the creative team can then curate and refine. AI accelerates execution of creative ideas and opens up aesthetic possibilities, yet the art direction – the guiding vision, taste and narrative coherence – remains very much a human endeavor.
June 2023
Project Type
Visual Research & Development / Experimental Art Direction
1–Overview
YTH is a self-initiated visual research and development (R&D) project that explores how AI can augment art direction, emotional storytelling, and material experimentation across fashion, product and image-making. Guided by my background in art direction, YTH combines tactile materials with abstract visual poetics — using generative AI (Midjourney & Flora) not just as a tool, but as a collaborator in crafting a sensorial world.
2–Creative Direction & Process
Hybrid Direction: What does it mean to art direct with AI instead of over it?
Narrative Core: YTH began with a story — a surreal ecosystem where textures feel, objects perform identity and reflections become symbols.
Material Emotion: How can textures (glass, fluff, chrome, skin) behave like characters?
Visual Research with AI: I used AI as an ideation partner to generate material studies (chrome typography, glass flora, soft ornamentation) to explore forms that blur physical and digital boundaries.
Curation & Styling: The most compelling AI outputs were selected, composited and refined using digital post-processing. This process applied prop styling principles and fashion sensibilities to abstract elements.
Narrative Form: Can non-literal imagery construct a storyline through visual rhythm?
Emotion & Sequence: I treated each image as a cinematic moment. Together, they form a visual arc — from fluid object metaphors to styled portraits and spatial illusions — evoking themes of identity, memory and transformation.
The LYTH logo itself became a central object of experimentation.
Starting as a 2D typographic sketch, I translated the form into 3D using sculptural modeling. This allowed me to treat the wordmark as a physical material study — refining it into a molten glass form that embodied the themes of fluidity, elegance and artificial tactility. The logo functions not just as identity, but as a recurring artifact throughout the world of YTH: a mark, a symbol, a surface.
3–AI-Based Visual Research
Using Midjourney and Flora, I generated hundreds of prompts based on mood, material, and form. Instead of referencing existing imagery, I treated the AI outputs as provocations — glimpses into possible material fictions.
“Plush chrome typography”,
“Liquid floral structures”,
“Futuristic softness cast in glass”
Each prompt returned unexpected, sometimes uncanny results. These were iterated, branched, and curated into a visual lexicon that defined YTH’s emotional atmosphere.
4–Curation and Concept Refinement
Next comes the curator’s eye — arguably the most crucial human contribution. I review the bounty of AI-generated material and select what aligns best with the narrative and emotional goals. This is where design principles and intuition guide my choices. I might pin up 20 AI images on a virtual board and notice, for example, that certain color palettes or forms keep recurring in ways that support the story. Those become keepers. Others, while visually striking, may stray thematically — so I set them aside.
It mirrors the traditional process of editing photographs or illustrations: now, I’m editing AI-generated visuals. At this stage, I might also combine elements from multiple outputs — a practice of compositing. For instance, if one image has the ideal background and another the perfect figure or object, I’ll bring them into Photoshop (or even AI-mixing tools) to merge them.
The curation phase often reveals gaps too. I ask myself: Do I have all the pieces I need to tell this story? Is something missing? Perhaps YTH already has several strong, product-like images but lacks a human element to ground the narrative — so I’ll return to Midjourney and generate a figure or mood that fills that gap.
This back-and-forth is iterative: generation, selection, refinement — then generation again, if needed. By the end of the process, the project begins to coalesce into a coherent visual direction — maybe a set of 10–15 hero images that together convey the desired aesthetic and narrative arc. These are still prototypes, but they serve as the blueprint for the final deliverables.
5–Conclusion
Finally, the project is ready to be output in the desired format, whether it’s a digital gallery, a printed lookbook, a video or a presentation for a client. I ensure technical quality (high resolution images, color profiles, etc.) and a consistent style across all deliverables. I share the work with stakeholders or test audiences and remain open to feedback. If something isn’t landing emotionally as intended, they can iterate – perhaps swapping an image out, or adjusting the pacing of the narrative. One of the advantages of having used AI in the process is flexibility: if a last-minute change is needed (say a different background or an alternate material), the team can return to the AI tool and generate new options relatively quickly, then integrate them. This agility is part of the model I’ve developed – it’s adaptable. After release, I might also reflect on the process: documenting what worked well with the AI collaboration and what challenges arose (e.g. dealing with an AI’s limits or ethical considerations). These learnings feed forward into the next project, continuously refining the hybrid methodology.
YTH is visual R&D as narrative poetry — a testbed for hybrid workflows, speculative styling and emotionally intelligent image-making. It offers a vision of how creatives can work alongside AI to develop new aesthetics that are sensual, strange and deeply human.
This project is both process and artifact: a design experiment that asks not just what we create with machines, but how we feel through them.