Rendered Real: The Noir Starr Podcast
"Rendered Real: The Noir Starr Podcast" dives into the intersection of high fashion, artificial intelligence, and authentic representation. Hosted by the visionary team behind Noir Starr Models, each episode explores how the digital modeling revolution is reshaping beauty standards, brand storytelling, and the future of talent.
Rendered Real: The Noir Starr Podcast
🎙️ Episode 35 — Prompt to Collection: The Evolution of AI Fashion Design
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
🎙️ Episode 35 — Prompt to Collection: The Evolution of AI Fashion Design
Fashion design is entering a new era—one where collections begin not with sketches, but with prompts. In this episode, we explore the shift toward a “prompt-to-collection” workflow, where generative AI acts as a creative partner, turning ideas into fully realized designs in seconds.
Designers can now visualize complex concepts instantly, using data-driven insights to replace traditional trend forecasting. Beyond creativity, AI is transforming production through 3D prototyping and precision pattern-making, reducing waste and accelerating development timelines.
As a result, the role of the designer is evolving—from hands-on creator to architect of ideas, guiding AI systems to execute their vision.
The future of fashion isn’t just designed.
It’s generated, refined, and scaled—at the speed of thought.
So imagine you are just casually scrolling through your social media feed, right? And you see this stunning model. Her name is Atano Lopez.
SPEAKER_00Oh yeah. The pink hair.
SPEAKER_01Exactly. The signature pink hair. She's 26, lives in Barcelona, and she's always posting about like fitness, gaming, her daily outfits.
SPEAKER_00And she has a massive following.
SPEAKER_01Well over 370,000 followers, hanging on her every single word. I mean, she pulls in up to$11,000 a month, doing real super lucrative brand deals for massive companies. Amazon, Razor, you name it.
SPEAKER_00Didn't she just launch a skincare line too?
SPEAKER_01Yeah. Vellum, real skincare line. But here's the crazy part. If you try to meet her in real life, you'll hit a brick wall. Because Itana has never taken a single breath of air.
SPEAKER_00Right. She completely does not exist.
SPEAKER_01Which is just wild to wrap your head around.
SPEAKER_00It really is. I mean, she is entirely synthetic. Every single strand of hair, the way the shadows fall on her face, how the clothes drape on her shoulders, it is literally all just pixels generated by a machine.
SPEAKER_01And Atana is really just the tip of the iceberg here. Because today, we are taking you on a deep dive into the literal fabric of the clothes you're wearing right now.
SPEAKER_00Because artificial intelligence, it's not just some back-end logistics tool anymore. It's not just tracking cardboard boxes in an Amazon warehouse.
SPEAKER_01Exactly. It is radically redefining the entire fashion industry from top to bottom. And to really grasp how insanely fast this is moving, we've pulled together a massive stack of sources for this deep dive, all totally current up to March 2026.
SPEAKER_00Yeah, we've got the big ones, the state of fashion 2026 report from McKinsey and Company.
SPEAKER_01Which is basically the industry Bible at this point.
SPEAKER_00Right. Plus, we're looking at deep technical profiles of the actual AI design software being used right now, and some really fascinating academic papers unpacking the hidden biases in these systems.
SPEAKER_01And I think the overarching theme across all of these sources, and tell me if you agree, is that the traditional kind of romantic era of fashion, like the lone designer sketching on paper, pinning muslin to a mannequin.
SPEAKER_00Yeah, that is fading incredibly fast. I mean, the industry was officially calling this the prompt to collection era.
SPEAKER_01Aaron Powell The Prompt to Collection era. Okay, so let's start there. Let's start at the very beginning of a garments life cycle, the ideation phase. Trevor Burrus, Jr.
SPEAKER_00Right. How a dress actually gets dreamed up in 2026.
SPEAKER_01Aaron Powell Because the sources show that designers aren't just, you know, sitting in front of sketch pads waiting for a muse to strike anymore. They're using generative AI platforms as just standard everyday studio tools. Trevor Burrus, Jr.
SPEAKER_00Yeah. Programs like Midjourney, Adobe Firefly, or specialized fashion platforms like newark.ai.
SPEAKER_01Trevor Burrus, and what used to take a whole team of illustrators weeks to do, like scouring vintage magazines, building these massive physical mood boards, sketching out 20 variations of a collar, that all happens in literally minutes now.
SPEAKER_00Aaron Powell It's instantaneous. A designer can just sit down and type a hyper-specific prompt. Something like uh a structured blazer inspired by 1950s brutalist architecture, rendered in translucent biosynthetic silk, shot on 35 millimeter film, and within seconds, boom, the screen fills with dozens of high fidelity concept images that look exactly like real photographs of a finished product.
SPEAKER_01Okay, but wait, let me push back on this a little bit. Let's unpack the mechanics of this. Because I'm looking at our notes on how these systems, generative adversarial networks, or jans, how they actually function.
SPEAKER_00Right, the underlying tech.
SPEAKER_01Yeah. If a jan just trains on millions of historical images and it predicts the next pixel based on what it's already seen, aren't we just going to end up in a loop of really boring average clothing?
SPEAKER_00That is a very valid concern.
SPEAKER_01Aaron Powell Because our sources even highlight this menswear collection by Acne Studios back in 2020. They generated it with AI, and fashion critics literally called it a math-crunched amalgam. So if the machine is just regurgitating the past, isn't it by definition backward looking?
SPEAKER_00It is. That's the core vulnerability of the technology, and honestly, it's a huge trap for lazy designers. But to understand why, you have to look at how a JAN actually works under the hood.
SPEAKER_01Okay, break it down for me.
SPEAKER_00It's essentially two different neural networks fighting each other. Imagine an art student trying to pass off a forged painting to a world-class art appraiser.
SPEAKER_01Okay, so the student is the AI.
SPEAKER_00Right. The student is the generator. They paint an image. The appraiser, who is the discriminator, looks at it and says, Nope, fake, the lighting is totally wrong. So the student tries again.
SPEAKER_01And they just keep going back and forth.
SPEAKER_00Exactly, millions of times per second until the student finally fools the appraiser.
SPEAKER_01Okay, so the system gets technically flawless at creating an image, but it doesn't actually understand what human culture considers cool or relevant.
SPEAKER_00Precisely. The machine has zero cultural context, it has no emotion, no intuition, it doesn't know that, say, an oversized shoulder pad feels rebellious right now because of broader socioeconomic anxiety.
SPEAKER_01It just knows what a shoulder pad looks like.
SPEAKER_00Right. And that is exactly where the human is irreplaceable. AI isn't replacing the designer, it's evolving the job title into something closer to a design architect or a prompt engineer.
SPEAKER_01Oh, that's an interesting way to phrase it.
SPEAKER_00Yeah. The human's taste, their curation, their ability to write that perfect prompt to guide the machine. That's the new differentiator. The mending gives you the variations, but the human provides the soul.
SPEAKER_01Which perfectly explains the Norma Kamali example in our research. I mean, she is a legendary designer. She's been shaping fashion for 57 years.
SPEAKER_00An absolute icon.
SPEAKER_01And in 2023, she actually went to MIT to take an applied generative AI course. But she didn't just log into some public AI tool like everyone else.
SPEAKER_00No, she did something brilliant. She developed a closed loop AI tool trained exclusively on her own 57-year archive.
SPEAKER_01And that closed loop distinction is huge, right?
SPEAKER_00Massive. Because if you use an open source AI, it's pulling from the entire internet, which just dilutes a specific designer's aesthetic with billions of random average images.
SPEAKER_01Right. It waters it down.
SPEAKER_00Exactly. But by closing the loop, Kamali created a walled garden. The AI can only reference her own past life's work. She actually calls it my Carl Lagerfeld.
SPEAKER_01Uh-huh. That's amazing.
SPEAKER_00Yeah. Referencing how the late designer had this famous reverence for archival inspiration. She's basically using the algorithm to endlessly recombine and propel her specific legacy forward.
SPEAKER_01Okay, so the AI helps the designer dream up the perfect collection, but predicting what you, the consumer, actually want to buy, that has also completely changed.
SPEAKER_00Aaron Powell Oh, trend forecasting is a whole new ballgame now.
SPEAKER_01Yeah, the sources mentioned this Paris-based company called Huratech. They use algorithms to scan millions of social media posts, street style images, runway data. And they successfully predicted the 2026 trends of like dotted prints and flat thong sandals months before they hit the mainstream.
SPEAKER_00Trevor Burrus, Jr. Which is wild accuracy.
SPEAKER_01Aaron Powell, but how does an algorithm actually tell the difference between a passing micro trend like, I don't know, three influencers wearing a weird shoe for a weekend at Coachella and a massive macro trend that's gonna dominate the entire year?
SPEAKER_00Aaron Powell Well, it comes down to analyzing the velocity and the network nodes of the data. HureTech isn't just counting hashtags. The algorithm uses computer vision to parse the actual pixels of an image. It can literally identify the specific cut of a flat song sandal, even if the caption just says, like, fun beach day.
SPEAKER_01Oh, wow. So it's not even reading the text.
SPEAKER_00Exactly. And then it measures how fast that specific visual data point jumps from avant-garde fashion accounts over to mainstream celebrity accounts, and finally down to everyday consumer posts.
SPEAKER_01It maps the contagion of an idea.
SPEAKER_00That's a great way to put it. And if you can map that contagion accurately, you completely change the risk profile for a brand.
SPEAKER_01Which brings us to the physical world. Because look, an AI can predict that a million people want a dotted print shirt. And an AI can design the absolute perfect version of it. But a digital JPEG doesn't keep you warm.
SPEAKER_00No, you still have to manufacture the physical thing.
SPEAKER_01Right. And actually making that shirt right now in 2026 is harder than ever. If you're listening to this deep dive right now and you're wearing a pair of jeans, the journey those jeans took to get to you was dictated by what the McKinsey report calls terrorist turbulence.
SPEAKER_00Yeah, the macroeconomic reality of 2026 is incredibly stark.
SPEAKER_01Tell me about it. Global trade maps are basically being redrawn from scratch. The report details how U.S. tariffs on imported apparel and footwear spiked massively.
SPEAKER_00Huge jumps. We saw them go from an average of about 13% all the way up to 54% in the spring of 2025.
SPEAKER_0154%.
SPEAKER_00I know. Before they finally settled around 36%. And for an industry where the U.S. imports nearly 89% of its apparel, that is a seismic shock to the system.
SPEAKER_01And those costs don't just magically disappear into the ether. Brands like Levi's, Nike, Ralph Lauren, they're being forced to raise retail prices just to protect their margins.
SPEAKER_00And supply chains are aggressively decoupling as a result.
SPEAKER_01Right. The data shows U.S. apparel imports from China drop by 30%, while imports from Cambodia surged by 42% purely because of more favorable trade terms.
SPEAKER_00It's all about dodging the friction.
SPEAKER_01Exactly. So if these tariffs are acting like this massive friction tax grinding the global supply chain to a halt, is AI essentially the digital WD-40 they're spraying into the gears just to keep the whole machine moving?
SPEAKER_00That is the perfect analogy. Honestly, it is a matter of sheer survival at this point.
SPEAKER_01Aaron Powell Really? Survival.
SPEAKER_00Absolutely. Because when your input costs and your shipping taxes skyrocket like that, you have to claw back savings somewhere else in the pipeline. Look at Levi's CEO, Michelle Gas.
SPEAKER_01Aaron Powell What are they doing over there?
SPEAKER_00Aaron Powell Well, according to the McKinsey report, she is rewiring the company end-to-end with AI to achieve hyper-accurate inventory management. They are aggressively pairing back less productive SKUs, meaning individual clothing styles.
SPEAKER_01Aaron Powell So making fewer options, but the right options.
SPEAKER_00Yes. The goal is to hit 85 to 95 percent accuracy in demand forecasting. You simply cannot afford to manufacture a pair of jeans that's just going to sit unsold in a warehouse gathering dust.
SPEAKER_01Aaron Powell And our sources also point to tools like style 3D AI being used to offset these massive costs. This software it takes out an initial 2D AI sketch we talked about earlier and generates a higher accurate 3D digital prototype.
SPEAKER_00Right.
SPEAKER_01But my question is, how is that different from just looking at a 3D rendering in like a video game?
SPEAKER_00Aaron Powell Oh, it's totally different because it is a mathematical physics engine, not just a visual model.
SPEAKER_01Aaron Powell What does that actually mean in practice?
SPEAKER_00It means the software calculates the actual physical properties of the fabric. It computes the mass of the digital thread, the friction coefficient of the weave, the exact tension required to sew a seam.
SPEAKER_01Wait, really? It gets that granular.
SPEAKER_00Oh yeah. So if you put a digital leather jacket on an avatar and make the avatar walk, the software simulates exactly how that specific weight of leather will crease and catch the virtual air compared to, say, a light chiffon dress.
SPEAKER_01Aaron Powell Well, so a brand doesn't have to manufacture a physical sample in Cambodia, ship it all the way to New York, realize the drape is wrong, redesign it, and then ship another sample across the globe.
SPEAKER_00Aaron Powell Exactly. You just manipulate the physics engine right there on your screen. It saves weeks of time and thousands of dollars per garment.
SPEAKER_01The digital WD40 in action.
SPEAKER_00Yeah. When margins are this compressed by global tariffs, squeezing efficiency out of the prototyping phase is one of the very few levers brands have left to pull.
SPEAKER_01Okay, so the garment is designed by an algorithm.
SPEAKER_00Yeah.
SPEAKER_01Its supply chain route is optimized by a mathematical model. Now, where does AI fit into my actual shopping part?
SPEAKER_00Oh, this is where it gets really fun.
SPEAKER_01Yeah, let's step onto the synthetic sales floor because this is where the industry pivots from just efficiency into something that feels like a full-blown sci-fi novel, which brings us right back to Atana Lopez.
SPEAKER_00The fake influencer making 11 grand a month?
SPEAKER_01Exactly. Why would a massive brand pay a computer-generated model to sell real physical clothes to real people?
SPEAKER_00To answer that, you have to look at why her creators built her in the first place. The creative agency behind Atana is a Spanish company called the Clueless. Right. And the founder, Ruben Cruz, stated very clearly that they were just tired of dealing with human influencers.
SPEAKER_01Because humans are messy.
SPEAKER_00Totally. Humans have egos, their schedules are unpredictable, they can get involved in personal scandals that damage a brand's reputation literally overnight.
SPEAKER_01So they just built a model who never sleeps, never complains, and is perfectly 100% controllable.
SPEAKER_00Yeah, it offers brands absolute safety and reliability.
SPEAKER_01But it's not just the models, the clothes themselves are also becoming synthetic. I'm looking at this company, D DressX, in our notes. They are a digital fashion company with over 100 million virtual assets.
SPEAKER_00100 million?
SPEAKER_01It's insane. They have an augmented reality app where you essentially try on digital clothes over your photos, and they partner with 4,000 luxury retailers to sell clothes that do not physically exist.
SPEAKER_00Just pixels.
SPEAKER_01Right. And wait, I'm stuck on this D-dress X thing. You're telling me people are paying real hard-earned money for a JPEG of a dress? I mean, why wouldn't I just use a free Snapchat filter? What is the actual economic value here? Are we just in the Emperor's new clothes for the digital age?
SPEAKER_00It sounds like it, right, but you have to think of it less like a temporary Snapchat filter and more like buying a high-end skin for a character in a video game. No. Except the character is you, and the video game is your professional LinkedIn profile, your Zoom calls, or your Instagram feed.
SPEAKER_01Oh, that makes sense.
SPEAKER_00We spend a massive portion of our lives looking at screens now. For many consumers, their digital presentation is just as vital to their social standing as their physical clothes. DSX provides digital assets that have verified scarcity. You are buying digital utility and identity signaling.
SPEAKER_01And the way you and I search for both physical and digital clothes is also undergoing this massive shift. The McKinsey report highlights a staggering statistic. Between July 2024 and July 2025, shopping-related searches on generative AI platforms skyrocketed by 4,700%.
SPEAKER_00Which is just an astronomical jump.
SPEAKER_01People aren't just typing blue suit into Google anymore. They are having full-blown conversations with AI chatbots to find what they want.
SPEAKER_00Which forces brands to completely rethink their marketing strategies from the ground up. I mean, for 20 years, the holy grail was SEO search engine optimization. Right.
SPEAKER_01Stuffing your website with keywords so Google ranks you first.
SPEAKER_00Exactly. But now brands have to shift to GEO, generative engine optimization.
SPEAKER_01Okay, how does GEO actually differ from standard SEO in practice, like for the person shopping?
SPEAKER_00Think of it this way: SEO is like putting a massive glowing sign outside your store that just says blue suit. It relies on exact keyword matching.
SPEAKER_01Very literal.
SPEAKER_00Right. GEO is entirely different. It requires giving the AI a comprehensive, deeply contextual dossier about that suit.
SPEAKER_01So treating the AI like an actual stylist.
SPEAKER_00Precisely. If a consumer asks an AI chatbot, hey, I'm going to a fall wedding in Tuscany, the dress code is smart casual, what should I wear? The brand needs their data format so the chatbot understands the exact shade of the suit, the breathability of the fabric, and its cultural appropriateness for an autumn event in Italy.
SPEAKER_01So the AI synthesizes context, it doesn't just match words.
SPEAKER_00Exactly. The AI acts as the personal shopper.
SPEAKER_01And if we project this forward, because our sources talk about this concept of agentic commerce, what does that mean for you, the listener, trying to buy that suit next year?
SPEAKER_00It means that in the very near future, you won't be doing the shopping at all.
SPEAKER_01Wait, what?
SPEAKER_00Seriously. An autonomous AI shopping agent will learn your exact style preferences, your precise body measurements, and your budget. And when you need that suit for the wedding, your personal AI agent will just go out into the digital marketplace, find the items, negotiate prices in milliseconds with the brand's own AI agents, and purchase the suit for you, bypassing traditional websites entirely.
SPEAKER_01My AI haggling with their AI while I just sit back and drink my morning coffee.
SPEAKER_00Yep. That's agentic commerce.
SPEAKER_01That is wild. But and this is a big but handing over our aesthetic choices, our global supply chains, and our consumer habits to these algorithms. Yeah. It means we really have to look at the structural seams that are holding all this together. Trevor Burrus, Jr.
SPEAKER_00The ethical implications are massive.
SPEAKER_01Yeah. The final section of our source stack involves these academic papers outlining the societal implications of this shift. And they're raising some serious flags, particularly around bias in the machine learning data.
SPEAKER_00Aaron Powell Right. The academic research focuses heavily on a concept called lookism and how these mathematical models perpetuate historical biases.
SPEAKER_01Aaron Powell Because we discussed earlier how a JAN trains on millions of historical fashion images.
SPEAKER_00Aaron Powell Exactly. But let's be honest, the history of the fashion industry is notoriously exclusionary.
SPEAKER_01Aaron Powell Highly exclusionary.
SPEAKER_00Trevor Burrus So if you feed an algorithm a massive data set where the vast majority of the models are tall, ultra-thin, and predominantly white, the machine doesn't understand the Kinklex social context behind that data.
SPEAKER_01Aaron Powell It just mathematically equates those specific physical traits with the actual concept of good design or beauty.
SPEAKER_00Aaron Powell Correct. Without deliberate manual intervention by engineers to diversify the training data, the AI just acts as a mirror and it reflects and amplifies the biases of the past at an unprecedented global scale.
SPEAKER_01And then there is the massive legal gray area of copyright law, which is a whole other mess.
SPEAKER_00Aaron Powell Oh, it's a minefield.
SPEAKER_01Let's say a prompt engineer types in design a tartan punk collection in the style of Vivian Westwood, right? And the AI spits out this brilliant new design that clearly learned its aesthetic from Westwood's actual archives. Exactly. Is it the person who wrote the prompt? Is it the tech company that built the AI? Or is it the estate of Viviane Westwood?
SPEAKER_00Aaron Ross Powell, Currently, the legal framework simply does not exist to answer that question definitively. We are operating in an absolute wild west of intellectual property right now.
SPEAKER_01And alongside all those legal questions, we have to look at the huge economic shifts in human labor.
SPEAKER_00The McKinsey numbers on this are sobering.
SPEAKER_01They warn that up to 30% of employee time across all industries could be automated by 2030. And in the fashion sector specifically, that means up to 40% of workers in developed countries, like pattern makers, technical designers, supply chain analysts, they will need to aggressively reskill or completely change roles as AI takes over those routine tasks.
SPEAKER_00It's a massive workforce transition.
SPEAKER_01But the tension that really stands out to me in these academic papers is the debate around sustainability. Because it is a complete duality.
SPEAKER_00It's the ultimate double-edged sword.
SPEAKER_01Yeah. On one hand, you have massive brands using AI demand forecasting to ensure they only manufacture exactly what will sell. They avoid overproducing millions of garments that would otherwise just end up rotting in a landfill.
SPEAKER_00Which is a massive win for reducing physical fabric waste.
SPEAKER_01Right. But the flip side of that equation is the hidden environmental cost of the technology itself. It feels a bit like taking a miracle diet pill that ends up causing heart problems, you know.
SPEAKER_00That's a great way to look at it.
SPEAKER_01Like we cure the disease of fabric waste, but AI makes the design and production cycle so fast and so cheap that it is supercharging fast fashion into ultra fast fashion.
SPEAKER_00And the computing power required to run all these generative models, the physics engines, those autonomous shopping agents, it all relies on massive sprawling data centers.
SPEAKER_01Which consume vast, staggering amounts of electricity and water for cooling. So we solve the immediate physical problem of fabric waste, but we are scaling a massive new problem of electrical waste and data center emissions.
SPEAKER_00It is the defining paradox of this current technological wave. Technology almost always solves one critical bottleneck while simultaneously scaling another one.
SPEAKER_01It's so true.
SPEAKER_00And the ultimate challenge for the fashion industry in 2026 isn't a debate about whether or not to use AI. I mean, that ship has sailed. The macroeconomic pressures have made it a mandatory business utility just to survive. Right. The real challenge is determining how to deploy it with rigorous human oversight. The industry really has to figure out how to use these algorithms to amplify human potential and solve real inefficiencies, rather than just exploiting the planet and the consumer's wallet faster than ever before.
SPEAKER_01Absolutely. So what does this all mean for you, the listener? Whether you are a high fashion aficionado tracking the runways in Paris, or you're just someone who buys basic cotton t-shirts in bulk, AI is now the invisible hand in your closet.
SPEAKER_00It touches every single thread.
SPEAKER_01It mathematically dictates what colors are manufactured. It calculates how much they need to cost to offset global terra friction, and it determines exactly how they are marketed to your specific digital profile through these conversational chatbots.
SPEAKER_00The boundary between human imagination and machine execution is practically gone.
SPEAKER_01We used to sculpt our clothes out of raw fabric, relying entirely on the physical touch of an artisan. Now we conjure them from massive lakes of data, but I want to leave you with a final thought to mull over.
SPEAKER_00Let's hear it.
SPEAKER_01As we plunge deeper into this era of agentic commerce, you know, a world where an AI algorithm predicts exactly what you want to wear, and then your autonomous AI agent goes out and buys it for you based on those predictions. Does personal style actually exist anymore? Or are we all slowly just becoming physical mannequins, walking around the real world, unconsciously displaying the algorithmic taste of the machines?