I've lived in NYC for the last decade+. Had many friends that had legitimate careers as models. Some lasted a year with just a few shoots and a self-funded trip to Paris fashion week before they quit or went into debt over it. Some have been going at it for more than a decade and you would recognize if you flipped through a fashion magazine even semi-regularly.
I also work in media, date someone in fashion and have knowledge of what actually gets paid to models.
While there are a rare few who find a real career out of it and rise out of the traps of shitty agencies and contracts, usually by branching out and establishing a self-brand, none of them that I know made something they can comfortably retire on just with what we would consider the job of a model.
There are certainly exceptions to this but in general, if you're a fashion brand, digital or print magazine offering any type of exposure, hiring a model is inexpensive. Rarely if ever livable wages. That goes for most of the people who work on sets or for fashion shows.
So with all of that preamble out of the way, what I am getting at is...
1. Coordinating an AI model, that has to wear these clothes, and that bracelet, and be on this location, or pictures with this lighting sounds complex and expensive when hiring a set to produce the real thing is a known quantity and cost virtually minimum wages.
2. There are still people who deeply care about the art of the whole thing and do most of their work for free to be supported in anyway to keep doing it. I am looking on the not so bright, bright side here but I'd like to think AI is little more of a thread than stock photography.
You know a lot more about this than me, but one potential counterpoint though is Ikea already using ~75% computer generated images in their catalogues. When the technology is mature enough, it provides a huge amount of flexibility compared to a photo shoot.
From my own attempts at hobbyist 3d modelling, computer generating a human with clothes is far harder than furniture. Looking through the computer generated images for IKEA almost all of them are hard body (solid objects) with very few soft objects (clothing, tablecloths, etc) and nothing that is alive. Creating realistic clothing renders is quite difficult, and it gets even more difficult when a lifelike human needs to be rendered as well.
Consider the incentives fast fashion vendors have, and it seems clear that completely AI generated product photography is inevitable.
It's expensive and slow to hire a photographer and a studio and a model, and you have to ship the clothes there ahead of the launch on your website to have them spend all day getting in and out of different outfits while stylists keep their hair tidy, then the photographs go to the art direction team to be photoshopped to match the site aesthetic...
If you can just get someone in the factory in China to snap a photograph of the latest batch of dresses and tops and skirts as they come off the sewing table, then you can just send them into a GAN, and have style-matched 'photographs' generated showing the clothes on a selection of different models, each of whose appearance is perfectly tailored to appeal to different market segments.
You can have high quality creative on your website and in the product feed to Google the same day, and start taking orders before the inventory starts piling up.
Then next week, you can do it again with the next set of designs.
> If you can just get someone in the factory in China to snap a photograph of the latest batch of dresses and tops and skirts as they come off the sewing table,
Just a simple photo of the garment will not tell you how it behaves on the body (how "malleable" it is, how it bends, wrinkles etc.) and how it interacts with the light.
Just ask any artist about the nuances of painting clothing materials - it's a big subject in its own right. I suspect that only shooting photos on a in-factory models, in various lighting conditions MIGHT be enough to train the AI.
Yep, but how is that informative for a particular garment that you want to visualize? You can't infer, from a corpus of images of random clothes being worn, how that particular shape and fabric behaves.
You don't need to have full computer generation to dis-employ a lot of models. Just making modeling a lot more productive would do that. For instance, you could have a "deepfake" approach that replaces the clothing a model is wearing. That would enable a small number of models and a pile of software to generate a hell of a lot of imagery.
From what I know about the industry, there's also different categories. I don't doubt that Target or Walmart could use digital models for generic clothing. I doubt that would work for a designer launching a new high end collection. Then there's the experiential side where designers like Oscar de La Renta will have events with live models for women to view their red carpet/wedding dresses. Potentially, software could help designers quickly test and iterate on new clothing ideas, but launch brands likely will launch campaigns around real people.
What I meant to say was not that the technology was ready today, rather that when it becomes ready it's likely to be used. At least it was in the case of IKEA when the technology was able to meet their needs.
The way modeling can work is that you take a photograph of a mannequin with the clothes upon it then digitally add in facial features or skin tone.
Even just being able to shift skin tone, and not show a face will work and will add to profitablily because people do want to know that clothing will work with their complexion
Ehh, this could backfire if it ends up in the uncanny valley realm. Consumers prefer images of models wearing clothing, because it shows the fitment. They may flock to brands that explicitly use real models.
IMO, modeling is one of those things that will never go away because of how cheap this labor is compared to hiring someone to make a software solution. Just like how it is cheaper to stock a fast food joint with a few minimum wage people worked to the bone than a sleek robot with a pricey service contract, even if the latter is sexier.
They were at 75% in 2014. They might be closer to 100% now. Ever since I learned about IKEA's CGI work, I always take more time to browse their catalogs.
While I agree with what you say, I think there's also subsets of modeling which can be cheaper with a computer which doesn't need wages, an agent, travel, or royalties. Also remember that models are human talent and human talent tends to come with costs which must be geographically local: makeup, photographer, scene, etc. Humans can only work so many hours a day, they can develop drug or eating habits, they can age. They can say indecorous things which will cause an outrage mob to want to boycott brands associated with them. All of these are costs or liabilities.
Granted, there may be a new generation of agents that specialize in AI models and royalties may still exist (with shrinking margins), but if nothing more, AI is likely to opt downward pressure on wages/jobs/contracts some of the non-minimum-wage models. Once it's bootstrapped, it will either become more appealing (for the reasons I mentioned above) or turn out to be complex and not worth the cost/risk. Only time and experimentation will tell.
> Human talent ... can say indecorous things which will cause an outrage mob to want to boycott brands associated with them.
I'm not envisioning a future in which some company creates a full persona for their AI models, and we get a full Tay[1] moment out of it, and then we've come full circle.
There are also the companies producing or providing the AI. Any entity capable of producing a convincing AI fashion substitute can do much more besides, some possibly not wearing well with public opinion.
Sure, your comment didn't look like it was trying to disagree as much as add more info, whether you realized I had a typo or not. I had just happened to read a comment I had mangled atrociously earlier because I did it on my phone and then I saw this comment, and the urge to correct myself here was overwhelming. :)
Also remember that models are human talent and human talent tends to come with costs which must be geographically local: makeup, photographer, scene, etc.
This sounds vaguely similar to initial discussions I remember with self-diving cars. It begins with "I look forward to the day we don't have fallible humans at the wheel" and end with the realization the AIs are going to be more fallible and in different, strange ways.
I mean, hypothetically, suppose you had a system that could constantly monitor world fashion trends, world clothing markets, the routes of hip and average people and so-forth.
In the case you might create a system that produced a variety of still and moving images that satisfied all the constraints that today's fashion industry satisfies just with a few written or spoken suggestions from executives. Then you'd eliminated no just models but a large chunk of the industry.
But let's look at what "AI" is now (and what it seems likely to be for a while without any "revolutionary" changes). What you have now is a way to extrapolate typical objects out of a stream of similar objects. Just GPT-3 does a great job create texts that sound vaguely right, you can create a vaguely plausible looking set of still and moving images of one or another "typical" model. Moreover, these extrapolations require constant training by professional much more highly paid than actual models now (as the GP notes).
Further, without being in the fashion industry, I'm pretty sure there's a lot more to a useful set of images than "looking about right". I suspect you could generate a model-image that would "work" with a kind of clothing (since both clothing and image can be trained). But generating a model-image that suits a given demographic, that expresses "what's becoming hip right now" and so-forth would be extremely hard.
It may not be impossible but it would require lots of high paid labor by AI engineers, defeating the entire purpose once the novelty wears off. And all this is to say that these "replace human activity" approaches wind-up with the problem of doing an "90%" of the activity right and then foundering on corner cases - like self-driving cars that are easy-yet-impossible AI tasks.
> I suspect you could generate a model-image that would "work" with a kind of clothing (since both clothing and image can be trained). But generating a model-image that suits a given demographic, that expresses "what's becoming hip right now" and so-forth would be extremely hard.
I mean, when I picture the problem, I imagine taking a single panorama-like photo capture with a smartphone, circling repeatedly around a human being whose likeness you want to use; and the phone using its barrage of sensors and ML cores to spit out a pre-rigged and textured high-poly 3D model of that person, that you can then drop into Blender and throw your clothing designs onto (i.e. the very same digitally-simulated designs that your designers prototyped with before getting the design made for real — presuming there was any amount of industrial design going into the object, which there certainly is for anything as complex as e.g. glasses frames, or a handbag.)
The pre-rigged 3D model output from such a body-scanning app would have a standardized rigging, such that 1. you'd know how your digital clothing items would interact with it before attaining the body-scan itself; and 2. allowing you to throw some posing "behaviour" scripts on it (that target said standardized rigging.) So this could all be parallelized.
Last step: pick a 3D-recreated environment, set the viewpoint camera and lighting, and snap screenshots at will. (This part doesn't need to be a science; you can just put a trained photographer in VR goggles, and have them circle around the digital model taking digital pictures with their field-of-view at time of trigger-press being the composed shot.)
The important part of this, from a cost perspective, is that you can then reuse this model for a combinatoric number of "shots", without ever paying the original body-scanned person again, or taking time to organize a new physical shoot with them. You can "re-shoot" them in localized advertisements for every target market you're launching the product in, all without needing to leave the room, let alone paying them to come back in. If you launch accessory products months later, you don't need to retain their talent; you have them "on file." Likewise if you need to dredge them up 10 years later for an anniversary "shoot."
> But generating a model-image that suits a given demographic, that expresses "what's becoming hip right now" and so-forth would be extremely hard
Just generate a spread of different looks, test them on different audiences, and measure engagement. Feed that back into the parameterization for generating the next set of images.
> And all this is to say that these "replace human activity" approaches wind-up with the problem of doing an "90%" of the activity right and then foundering on corner cases - like self-driving cars that are easy-yet-impossible AI tasks.
I don't think the boycott mob is a unique problem to the human. Could just as easily end up with an anti-AI boycott mob. Neither bet is fully safe on that front.
The issue is for companies that just want to sell clothes, without also being forced into a wide variety of shallow, poorly thought out political stances that could blow up in the companies face. Note just how much of this article is the 'model' talking about what is basically politics. They're half model, half politician. This is 100% liability to most of their employers, who ultimately just need a good looking mannequin to wear some clothes or pose seductively for advertising.
A GAN generated face textured onto a convincing 3D model means you can just keep clicking or adjusting until you got a truly beautiful model, without needing to pay extra for it, and you know your exposure is capped in entirely predictable ways. This model won't unexpectedly cause trouble for you, or demand you take a particular stance you may not agree with in order to employ them.
I don't see the average consumer caring whether the model for their perfume commercial is flesh and blood or CGI. They already don't care what the quality of life is like for the test animals for those same cosmetics. Plus, it's easy enough to create fake personal social media profiles to pretend like the persona has a real history.
It seems like you're looking at the core of the job as it exists now, but perhaps the threat will be on the periphery? New technology tends to automate tasks, not jobs, but that can change the jobs.
Stock photography probably does have some effect, for some websites where they might have hired a model. Suppose stock photography gets better, more flexible? What could a more ambitious stock photography company do to help clothing retailers find a different way to sell clothes?
It depends on the kind of shoot. A picture for a catalog is going to be much cheaper than a picture for an ad. The reason isn't just the model's time, it's all the other work that goes into it. A picture for an ad, all things considered, can be surprisingly expensive. So the catalog images are easier to do digitally and the ad images have a larger incentive to make cheaper. If you can replace enough of the pipeline, you can save significantly. The model is just a part of that, but they'd still lose their job. Even the simple catalog work where you might digitally change that solid red t-shirt to blue and green and orange saves time and means fewer models are needed, shrinking the job market.
It's hard to imagine AI/virtual models being able to satisfy the high-end fashion industry's love for traditional pomp and pageantry. But I'd have to guess that an unlimited, cheap supply of perfect and customizable human models will unavoidably have a massive impact on the many non-A-list models who make a living doing photoshoots for unbranded campaigns, especially models who are currently used to model clothing for online sellers.
While there will likely always be added commercial value for (human) celebrity campaigns – e.g. Kanye and Gap, Jennifer Lawrence and Dior – I'm not sure how Old Navy/Banana Republic/Uniqlo/etc. would suffer much at all by having digital models for their website and in-store photography.
There are different kinds of modelling. The old paper catalog - now website - modelling has already been replaced. You can see the results all over Amazon and various merch shops. Clothes and other objects have colours and textures shopped in effortlessly.
Is that AI-able? Not yet. Making edited images look seamless is still a moderately skilled job, and AI is still struggling with basic object recognition, never mind the semantics of object presentation.
It might be possible one day, but not for a good few years.
High end modelling is about celebrity, and that's not going to be replaced any time soon.
Likewise for high end fashion photography. You can't hand something like Nick Knight's work over to an AI, because no AI has the creativity or imagination needed to make images that look like that, and engage the viewer like that.
It might be possible in principle to automate some of the more obvious fashion cliches - intensely aesthetic people with cheek bones in a variety of exotic locations - but it's harder than it looks, and the quality of manual production values will make it very hard for AI efforts to cross uncanny valley without getting stuck in it.
Attempts will also suffer from the CGI problem, where CGI turned out to be more expensive than modelling for most movies. And the results end up looking plastic and rather soulless no matter how much detail they have.
As I understand it, actor performance quality wasn't the main driver of The Mandalorian's live CGI sets. It was lighting and reflections. When your main character's head is essentially a chrome ball, green screens really aren't going to cut it. They needed believable reflections and lighting and the live set gave them that.
There are and always will be a lot of people willing to do art for free or even at great personal cost. I've had friends who modeled niche clothes (i.e. corsets) for the clothing producer, in exchange for a discount on the clothes (not even free clothes). I would go so far as to say that a lot of the most interesting art out there happens at break-even or negative valuations.
That said, I don't think you're right to place big companies that provide most of the media presence of fashion scenes in that category. Big clothing companies care about selling clothes, not about art, and they'll follow the cheapest, most effective way to do that. A marketing scheme based around supporting artists might happen, but if it happens it will be because market research says it plays well with target demographics, not because of some sense of charity. And it will likely be a token gesture, not a core strategy.
Look at what has already happened in fashion in the past: nods to fat shaming have been laughably tiny "plus sized" models, nods to race issues have been light-skinned black women with primarily European features, nods to skin not being perfect have been un-photo-shopped pictures of women who, from what I can tell, have perfect skin to begin with. And the vast majority of the time the gigantic Broadway/Lafayette billboard is a slender white photoshopped woman.
The cost of doing this stuff with AI is only going down. Why would you pay a whole photo crew and model when you can send a few low-rez photos of the clothes to a team in Bangalore and get back a video of a "model" with exactly the body specifications you request, doing exactly what you want, for $200?
Interesting. In some way, in hindsight, I believe the internet might also be responsible for the downfall of the profession. Remember the "Top Models" of the 90s and that this is less common today
In some way, the demand for modelling has probably gone up, but it's more long tailed. The internet also allowed for more "democracy" in this area and less gatekeeping
The different tastes and long-tailed nature probably contributed to less emphasis on "top models"/attention being focused on a sole person and/or mainstream beauty standards
I don't know nothing about all this. How much of work that involves models has an actual creative process behind it vs. having someone pose in front of a white background with a product for a catalog?
The latter seems to be bound to be replaced by AI eventually. I could see something similar happen like to orchestral music for movies and games where since years only few players, especially soloists, are recorded live and the rest is entierly made up by virtual instruments good enough to trick most people into thinking there is an full orchestra.
Think we are going to have real models for the magazine covers and expensive ads for a long time. But for e.g. online clothes shopping, to be honest, I would prefer to be able to switch out and modify the models to something closer to my body than what they usually are.
I've still never heard a virtual trumpet that sounds like a real trumpet. Not even close. Each trumpet player has his own tone or "lip" that is unique. I can tell who is playing just by a few notes.
Solo I don't think there are many virtual instruments that get even close to the real ones. But grouped together, mixed with real instruments, it gets a lot harder for the untrained ear. Especially in a context like a movie where the music is not the primary focus and often much less nuanced.
Virtual models gives you a lot of flexibility.
For example, you can change the skin color of the models depending on the country of the IP address.
You can even tailor the body types of the models for every customer to maximise the sales.
I wonder if, after a torrid love affair with AI models, we'll have to return to human models simply because nobody could resist the urge to tweak the AI model parameters beyond what is sensible and useful over time, as AI modeling users compete to stand out that little bit more. Convenience of flexibility could become a problem, in the long run. Not that models haven't had their own problems with trying to conform to whatever the industry's standards of beauty is this month, but having to still be living human beings has at least kept them on Planet Earth to some extent.
I think the problem is that models will become more and more "super-human" and computers will help with that. You can already see this in the amount of Photoshopping in ad photography. And for example in imagery aimed at children (e.g. the unrealistically big eyes of the Frozen characters). People want eye-candy and it doesn't necessarily have to be realistic.
At some point it will just be simpler and cheaper to replace Photoshop and in fact the entire photography/imaging pipeline by some AI.
If you can put your clothes on a virtual model in photoshop -and be done right there - that will be it.
I think this will start to happen in the next 5 years.
First for the 1/2 of fashion that is low-end and it will look a little off - but as colour and lighting and sets improve, it will make its way into other brands.
I should add: fashion and most creative industries are extremely cheap and cost conscious. Making that shirt for $19.99 entails cutting every corner possible. Every little bit of fabric etc. optimised to save money.
Once they can start generated images 'for free' they will, and it may not even be the unit expense, it may be the operational expense.
Chicago Tribune now sends reporters out with iPhones instead of having staff photographers.
Our food supply is full of filler and garbage ingredients.
Fashion brands are constantly dying, where there is a way to lower costs, it will happen.
given that the bespoke companies (Inditex, H&M) are regularly selling off half their stock at 50% off and meanwhile looking into insolvency every few years, I think you overestimate their execution. Producing a shirt probably costs 40cts (if at all) at enormous costs to communities in the disenfranchised parts of the world.
I doubt that the "brands" do any "finegrained" optimization, it's "make shirt or have no food"; the actual subcontractors probably do, but not in any rigorous way except introducing slavery.
They are viscously price-sensitive industries, always flirting with commercial problems, running on very thin margins with ugly working capital requirements.
These are exactly the kinds of companies that will use nearly-free AI instead of real models.
And FYI, they are excessively optimised for efficiency, more so than most industries. When an item doesn't sell, it goes on clearance right away. They adapt faster than any other industry to trends, down to the micro level. 'Home Base' knows exactly what is going on in every store and everything they make, cost & labour is a key consideration: 'this kind of seam -> this kind of cost' , 'this pattern doesn't fit well onto sheets of material leading to XYZ amount of waste, implies ABC extra cost'. Their online data is mapped to local inventories etc..
That's the only way 'fast fashion' exists, and it's a function of their scale, reach, adaptability etc.. In many ways, they are 'exemplary'.
> hiring a set to produce the real thing is a known quantity and cost virtually minimum wages.
The interesting question is whether the cost of the technology solution can be brought down below the cost of the human solution. If it can, then it's not a question of "if" the humans will be replaced but "when". I don't know enough about the cost structure to give an answer.
This. Modeling and photography are low skilled jobs, with minimal pay and long hours. Over years system became ruthlessly efficient to extract value from people.
Good luck replacing that with expensive AI developers to produce fake stuff.
I have no knowledge of the skill involved in (real life) modelling, but I do know the skill involved in professional photography is a lot higher than I fully understand.
For most of us, photography is just point the phone and tap the screen, without really giving much thought to lighting (colour/s, fill/spot combinations), scene composition, lenses, and probably a lot of things I don’t even have names for given the stuff I’ve listed is stuff I only know about from 3D modelling.
And conversely, the end users of a future AI synth of a model won’t be paying directly for expensive AI developers, any more than the average visitor of thispersondoesnotexist.com
>I have no knowledge of the skill involved in (real life) modelling, but I do know the skill involved in professional photography is a lot higher than I fully understand.
Yes, but there's no shortage of people who know all the involved stuff...
High-skill as in "you need to know lotsa stuff", yes.
High-skills as in "the skills are rare, and require a special degree or years or training", no.
They're not that rare (there's an overabundance of both skilled and non-skilled photographers), and they're not that hard to pick up (to the point that 18 year olds can know all there is to it with a little determination and practice).
Or let's just say that "high skill" is relative, and being a pro photographer is hardly like being a pro coder or a surgeon...
> there’s an overabundance of ... skilled photographers
I’m really not sure this is the case. I know more about cinematography than photography (I direct) but here in China a decent cinematographer can charge for a day what some workers might earn in a year. And you can tell the difference between their work and someone cheaper. That would suggest to me their skills are rare.
> and being a pro photographer is hardly like being a pro coder
This is where I strongly disagree. Becoming good at composition, setting up shots, etc is an art and can take a lifetime to perfect. No less high skill than programming.
High-skill usually implies that some sort of specialized training or schooling is required. Working an espresso bar is also a delicate skill, but no one calls baristas high-skilled workers.
Simple. The model can be the last part of a production pipeline in meatspace that gets replaced. There's a lot of post production that happens after a model shoot.
Render the clothes on the model before production to test demand. Render the model in the outfit the customer has in their cart right now.
I've lived in NYC for the last decade+. Had many friends that had legitimate careers as models. Some lasted a year with just a few shoots and a self-funded trip to Paris fashion week before they quit or went into debt over it. Some have been going at it for more than a decade and you would recognize if you flipped through a fashion magazine even semi-regularly.
I also work in media, date someone in fashion and have knowledge of what actually gets paid to models.
While there are a rare few who find a real career out of it and rise out of the traps of shitty agencies and contracts, usually by branching out and establishing a self-brand, none of them that I know made something they can comfortably retire on just with what we would consider the job of a model.
There are certainly exceptions to this but in general, if you're a fashion brand, digital or print magazine offering any type of exposure, hiring a model is inexpensive. Rarely if ever livable wages. That goes for most of the people who work on sets or for fashion shows.
So with all of that preamble out of the way, what I am getting at is...
1. Coordinating an AI model, that has to wear these clothes, and that bracelet, and be on this location, or pictures with this lighting sounds complex and expensive when hiring a set to produce the real thing is a known quantity and cost virtually minimum wages. 2. There are still people who deeply care about the art of the whole thing and do most of their work for free to be supported in anyway to keep doing it. I am looking on the not so bright, bright side here but I'd like to think AI is little more of a thread than stock photography.