I am a model and I know that artificial intelligence will take my job
vogue.com
vogue.com
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.
Good luck replacing that with expensive AI developers to produce fake stuff.
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
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...
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.
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.
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.
https://web.archive.org/web/20141230115206/http://www.cgsoci...
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.
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.
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
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.
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.
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.
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.
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'.
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?
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.
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.
The Mandalorian begs to differ.
The problem was that the actor couldn't see the CGI in real-time.
Once they built full wall displays so the actors could see what they were acting to, everything improved quite dramatically.
That will be covered by the top 100 models of the time. Those are the ones that have millions of IG followers.
The next tiers down will absolutely be replaced. The model interviewed about it has the exact same opinion because she actually lives that industry.
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
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.
If you're a big outfit looking to control risks, it might be very tempting.
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'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.
I'm reminded of the (aweful) film Simone (https://en.wikipedia.org/wiki/Simone_(2002_film)), and a Ben Bova novel (possibly Starcrossed) in which CGI gradually replaces human actors.
Lots of typos for me today... :/
Am also agreeing with you and extending the notion. AI is by no-means outrage-proof.
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 think this is the crux of the issue.
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."
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.
At some point it will just be simpler and cheaper to replace Photoshop and in fact the entire photography/imaging pipeline by some AI.
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.
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?
Save the condescension, this wasn't written by an 'IT type' or published on a tech journal.
All industries that got disrupted by more modern technology came up with arguments similar to the ones you brought up - bookstores vs amazon, brick-and-mortar stores vs ecommerce, face-to face meetings vs video calls, film vs digital, newspapers vs internet.
There will always be demand for high end fashion, but eventually it'll get relegated to a niche.
You only need to look around this very thread to see many indignant, condescending comments towards models by those who refuse to try to understand the value they bring to the creative process. It's much easier to condemn something than to try to wrap your head around a totally different social milieu.
Humans thrive on real in-person contact, but you wouldn't know it if you looked at how we acted.
Power looms certainly drove out hand looms, displacing many artesians that supplied most clothing in the 18/19th century. Suddenly clothes were cheap because of a new technology! But does that mean there's no market for specializing in the higher end clothing that requires special attention and detail? No! Instead, the market tends to bifurcate into the mass consumption market and higher end artesian market (I certainly know many people who do like the higher end artesian products!).
I think we need to worry less about if we have X or Y technology that will disrupt a working class of people, and instead focus on building up a more robust welfare state to allow these people to have a meaningful place in society.
We still have horses today. We still have coal miners. We'll still have human models. There'll just be less of them.
Think of all the teams of bookkeepers (yes, actual people who penciled numbers in books) who were obsoleted by Excel being able to let a store owner do a calculation/scenario by himself that would take accountants a week to do.
Think of all the secretaries whose work disappeared (or were no longer needed in proportion to the growing economy) as soon as personal calendar software and meeting invites became common.
Graphic designers / publication layout experts you would pay because you didn't have desktop publishing software.
There are more jobs lost silently to these kinds of developments than any factory being shut down dramatically. (for the US at least)
People think that economics is a zero-sum game, but the endless drive for efficiency and productivity is what makes our world possible and lifted billions out of abject poverty. It is the opposite of zero-sum.
Software empowers people.
Thankfully, as you point out, it often does the exact opposite and makes everyone richer! But not always, not reliably, certainly not as a fundamental guarantee. "The Wedge" plot illustrates this fickleness, where about 40 years ago the American story switched from "a rising tide floats all boats" to "rich get richer, poor get poorer." The economy kept growing, but the overwhelming majority of people not only did not manage to capture a share of the new growth, they did not even manage to hold on to what they had. Yes, those on top scooped up more money than those on the bottom lost -- but how is that supposed to be comforting?
It frustrates me when certain elements preach the prosperity gospel while framing it as a matter of fact rather than self-serving faith. I personally have faith that we'll eventually figure out a compromise, but I don't believe that denying blatant trends helps us get there.
But that never happens. Ten million multiplied by twenty million is 200 trillion. You can't make 200 trillion dollars by taking a dollar away from a million people. Nobody has ever made 200 trillion dollars from anything, but if they did, it makes no sense to think there's some way it could be done by impoverishing a million people who have almost nothing. It sounds as illogical as the Matrix use of humans as batteries.
Trying to put it in concrete terms - Jeff Bezos has maybe $180B. One ten-millionth is $18K. So the scenario is, suppose Bezos got his $180B by taking $1800 each from a million people, who relied on that capital to live, and somehow multiplied it by ten.
Since nearly everyone rich has less than Bezos, anything that happens a fair amount would involve smaller numbers.
It's not as absurd as making $200 trillion, but it still doesn't sound to me like a thing that happens to the extent that it says something about economics or utilitarianism or whatever. It seems like a contrived trolley problem to me.
Or trying to do multiple scans and then bagging them all in one swoop.
And god help me if I try to scale with 2 more hands.
If you care about accessibility and not being ageist, they are terrible for people with disabilities or the elderly. You will almost see no old or disabled person using a self checkout line.
As a show of more anecdotal evidence, a recent large grocery store chain in my large populated city of 2+ million people experiment with going self-checkout only failed so bad (lost so many customers and people were complaining), they hired cashiers again to basically scan people's groceries for them at the self-checkout line. Now they are stuck with the worst of both worlds.
You dutifully wander through the aisles gathering your goods, You slide your card, operate the pin pad, carry your own goods to your car. These are all goal directed activities you completed for YOURSELF in order that you could consume or use the products you have so acquired. There is no fundamental difference between scanning your own goods and sliding a card.
You aren't paying for someone to scan eggs and put it in a little bag you are paying for someone to manage every step between where the hen laid the egg and making it conveniently available on a shelf 1/2 a mile from where you live.
You aren't getting paid for doing it yourself like you aren't getting paid for carrying your own goods to your car instead you are benefiting from a price point enabled by the degree of automation and self service that the store engages in. Its ironic that people simultaneously flock to stores that have even slightly lower prices while complaining about lack of help. Simultaneously driving and bemoaning the same trend.
Your anecdote about bringing back the cashiers for the worst of both worlds sounds like a buggy whip manufacturer gleefully cackling at the unreliability of early cars. I believe we both know how THAT turned out. Given that Walmart was doing inventory on all its socks by walking past the socks with a wireless reader 10 years ago I'm pretty sure even your can of baked beans will have a chip in the label before long and your self checkout experience will be literally consist of solely being asked to pay for the goods in your cart. At this point paying an entire body to baby site each transaction would be wasteful and silly as 99% of them will consist of you touching a button on your phone or on store hardware to pay.
You say that self checkout is "ageist" in an era where even people turning 60 years old today probably saw a computer by the time they were 30. In 10 years this will be true of our 70 year olds. Are we just supposed to pretend that people who didn't have a phone shoved in their hand at 5 can't learn? That would seem in itself to be ageist. The reality is that old machines sucked pretty badly and older people don't like change and have taken their impression from older machines. This isn't the same thing as being incapable. For those that truly do have difficulties it ought to be sufficient to have staff on hand to help.
There is fundamentally no difference between waiting in line 2 minutes patiently in line and standing at a self checkout while the attendant helps others there for the same duration but people don't seem to react the same at all. In fact properly regarded what having 4 self checkouts with one attendant instead of 1 cashier with one computer is the probability of waiting far less.
Would you rather wait behind 3 other people or would you rather checkout out immediately and wait 30 seconds if you need help with the machine?
Get better self-checkouts. The ones that try to simulate how a traditional checkout works are stupid, the same way a mechanical messenger pigeon would be stupid but email is pretty good.
A friend of mine has several serious problems. He much prefers the self-checkout. No human interactions, which means no need to try to figure out what the other person thought they were communicating.
It was much better during the worst of lockdown too. I go in, I pick up items I want, I scan them and place them in my backpack, I go to the checkout, I hold the same device I used to scan items near the checkout, it acknowledges that I agreed to pay for the items I scanned, I put the backpack onto my back and I walk out of the store. Minimal contact, no human interaction, very low risk. Nice.
A digger (machine) replaced a lot of workers with shovels, but in the long-term it has clearly been good for society.
I think one way to solve it is creating institutions that invest in the job cutting technology on behalf of workers. This way workers pick up the productivity gain and not the businesses that employ the technology.
It would be a massive shake up and would seem very seizing the means of production in a indirect way though.
That night he picked up a manual and learned a little BASIC and put together a program to do some calculations for manufacturing bridge spans. It would normally take 3 guys 2 days double checking and redoing the precise calculations but the Apple II took minutes. Now 3 guys were free to do other things and a bottle neck was removed. The company could take on more work and the boss was pleased. “Take those things out of the trash!”
What the boss really didn’t understand was the software you needed to buy to make the computer useful.
And that's the conundrum we face with automation where we remove bottlenecks, and we can't retask the capacity.
I would almost change it to say "what the boss didn't really understand was how to use the computers to make my employees more productive". I think the specifics in how you automate is incredibly important, and there are many cases of how it can fail. 737 Max MCAS system is a prime example of how ramming automation through without consideration of "human factors" leads to massive failure.
| Think of all the secretaries whose work disappeared (or were no longer needed in proportion to the growing economy) as soon as personal calendar software and meeting invites became common. And hence forth came executive assistants, who could focus on more important features of their job such as managing a calendar rather than retying memos!
| Graphic designers / publication layout experts you would pay because you didn't have desktop publishing software. And now graphic designers can produce amazing movies, the complexity of which would dumbfound animators from the 1930s!
In almost all your cases, the productivity of these positions has grown, enabling more efficient use of their time and resources. Sure, it's required to know how to use Excel, Outlook, and Creative Studio to be productive in these newer jobs, but they it's precisely because we have integrated these tools into our workforce that we can be so productive. I see an analogy to asking "what are radiologists going to do when the AI comes"? Sure, maybe the older radiologists who don't use AI tools may be outdated and either learn to use newer tools or retire, but radiologists are not fundamentally going away. And other positions that may outright be antiquated, it's a moral imperative to create a robust welfare and career focused educational system to ease transitions pains.
However, decades later, there are still secretaries who schedule meetings and manage electronic calendars, bookkeepers who type numbers into excel spreadsheets, graphic designers and publication layout experts who are versed in professional publishing software, and even accountants.
The industry continues to be centered around still photographs—generally for the average campaign 90% of the budget/crew will go to the photographs, and video will be thrown in as an afterthought, even though it is an order of magnitude more difficult to create—and nearly exclusively those stills will be experienced on a computer that is told to show that same frame 60 times every second forever.
My clients are just barely starting to understand how video works. To try to get them to wade into 3D—and not just as a splashy one-off tool for attention, but for the actual day to day creation of hundreds of e-comm images/season—I don’t see this happening for a long time.
If they are very familiar with still photographs and (I assume) can somewhat predict how still photographs will be perceived by the market, what is the incentive to switch to something new?
As a consumer, my guess would be that a video or 3D display would not create a huge spike in revenue. In fact, if done poorly I could even see it having the opposite effect.
So what is the incentive to invest time and money into switching to something new and risky?
-------
CGI models however seem to be a different story. The cost saving aspect is clear cut and I as the consumer likely won't even realize anything has changed.
On one hand, it makes total sense to ask if embracing 3D stuff, or pushing (to my mind) a more appropriate use of the digital mediums in which we create and experience most things will lead to spike in revenue:
If you do it poorly (read: solution looking for a problem) I wouldn’t expect that to make much of a leap in any real metrics—and if companies are trying to pass off images on the wrong side of the uncanny valley that’d be more likely to hurt than help.
But it’s the Art that actually sells the “lifestyle” (read: clothes), and if you can create a gobsmacking incredible experience that makes people feel things you will absolutely see that in metrics and earned media and attention...
There are so many interesting technologies that are widely accessible today that fashion companies aren’t embracing because 1) they don’t know to look for them and 2) they don’t understand how they work. Small example: I absolutely blew a (publicaly-traded) client’s mind showing them a projection mapping concept... 2 years ago, well after the tools made it a 15 minute job they could have gotten the savvy intern to execute.
Ten hand-picked photographs will probably look better than the whole video they were picked from.
And to your point the funny thing about fashion and narrative-style film/video is that when it’s a single frame of a skinny lady making a contorted pose in the middle of a crazy scene, you’ll accept that as given in a single image, but suddenly when you actually have to flesh out the world she’s in and try to create an implicit narrative around it to keep the viewer interested (instead of just a vague moody simulacrum of depth) it falls apart; when fashion people talk about “story” they are usually referencing the relationship between particular garments and poses between group still images, without any regard for what that word means in the larger narrative sense re: it being the key to human’s communication and attention.
I think there’s more an opportunity to rethink the problem stills are solving, and if our current solutions are still the most interesting and medium-appropriate ways to address those.
Like say we’re talking functional e-commerce imagery to sell you a specific sweater. What’s the best way to communicate the weight and flow of that sweater: through a 1/125th of a second of it puffed up to show its shape, or seeing how it actually flows and how the weight responds to manipulation in real time. Why deprive ourselves of all of that rich lighting and movement and color dynamics information our brains use to understand what we’re seeing, just because traditionally we’ve used ink smeared on wood pulp.
See Marvelous Designer[1] and CLO[2]. These are CAD programs for designing clothes. They make both a 3D model for viewing and patterns for cutting and sewing. When design moves to CAD, the designer already has a 3D model before the clothing is made. So, for catalog photos, there's no need for human models.
Mostly. Those two companies need better hair shaders.
https://youtu.be/2IZfSr891bE (warning, nudity)
Two additional scenes stand out:
1. The protagonist watches a prototype perfume commercial with eye-tracking glasses, and the computer ends up superimposing the closing logo over the part he watched the most often (this being 1981, I'm sure you can imagine...)
2. The implication that the computer can determine the 'perfect' poses and actions for optimal viewer response ("Not enough body twist according to the computer"), and the physical model having to contort herself to fit the ideal (you can see a few seconds of this at 0:39 in the trailer - https://youtu.be/yoT-r1slAZ4)
Models are cheap, but the overhead of the process is expensive. Hiring a photographer, lighting person, studio space, model, clothes and backdrop; coordinating with relatively high-paid internal stakeholders (execs, designers, etc.); and developing/touching up photos after... Big processes add up. There's a need.
At the low and medium end, this could totally replace the shoot process. Presumably, designers would have a basic version of the software in their standard toolkit (you can see it in a catalog before it's shot - talk about sales!), so the marginal cost would be 0. There's no differentiator - no friction.
If the software's output is comparable to a shoot for a department store, the there's a real solution.
Why would I ever bother with a physical shoot?
There are untold small and big improvements like that that we just take for granted. GDP per capita roughly doubles in 20 years. That's the combination of all of these small and big improvements added up on a societal scale. Before industrialization it could take over 1000 years to see a similar level of improvement in the life of an average person.
My grandparents had no running water. They would wash in a sauna with water from a pond or well. Famines were common at that time. People still mostly used horses for transport. Roads were not paved. Clothing was mostly self-made. Televisions didn't even exist yet. Radios were for well-off families. Compare that to today in a developed country.
This rampant non-stop technological advancement is what's making life better. It's just hard to notice if you don't think about it.
https://www.cnbc.com/2020/07/10/looming-evictions-may-soon-m...
You caan't live in a video game. You cannot eat an iPhone. Maslow's Hierarchy still rests on its base, and tens of millions within the US and billions worldwide live precarious existences.
Living on a knife's edge, at all times, is not tenable.
See:
https://old.reddit.com/r/dredmorbius/comments/2vwfb6/maslows...
https://old.reddit.com/r/dredmorbius/comments/3ey7d1/maslows...
Based on the sorts of 'unplugging' trends that get picked up with some regularity, I'm not unique in this experience.
There is so much that's really cool in tech, but that feeling of discovery and wonder is a dopamine high. The junk food of happiness. It doesn't sustain, so you either have to let it go at some point, or get stuck in a loop of novelty seeking that doesn't end until you're just too old and tired to keep doing it. And like any addiction, the people who don't chose abstinence feel existentially threatened by those who do, and react as if being personally attacked. Meanwhile I'm sort of stuck in the middle because I don't think either works as well as moderation, which both sides hate because 'you people' won't pick a side. Novelty should be novel.
I keep waiting for the West to repeat the experiments of the '60s, complete with zen monasteries (now with 85% less sexual harassment!) and Hare Krishna robes everywhere.
At the risk of quoting a pervert: Everything is awesome and nobody is happy.
Replacing this with all digital models and clothes would be a big cost reduction.
However, it is still relatively hard to render photo-realistic faces and there's still a long way until all clothes are available as 3D models with realistic simulation of fabrics etc.
But there are already solutions being used today that achieve some of the benefits without using completely generated content.
Looklet[1] provides a system where each garment is shot individually on a mannequin. This is done by a couple of operators in a custom studio, typically placed in a warehouse or similar where samples are received. The images are then combined with other garment images and previously shot images of models to produce photo-realistic catalog images without the need for a traditional photo shoot. The web page has sample images and a list of retailers using this technology.
Take a look at e.g. Saks Off 5th's[2] catalog and see if you can spot the images that have been produced in this way.
Model is just a top of the pyramid which is being eaten by software.
One can see though that that may also lead to small tech-advanced (3d printing/etc.) object "materialization" shops popping up close to consumer. While you're running your morning run and having breakfast, the outfit chosen upon waking up (based on looking at weather and your own "feel like") from a design collection just posted couple days ago (and which you can preview online as fitted right onto you instead of a model - it may look good on a model and not on you and vice versa) is getting "materialized" and delivered right to your door (and your previous ones which you don't need/want anymore are collected for recycling, refurbishing, donation, etc.).
1. How long before someone plugs in a GPT-3 backed chatbot to handle the comments for these virtual models? Eventually, AI powered voice synthesis, lip-sync and animation (helped by a kinematics model) will handle basic animation, to allow real-time chat with a "virtual model" who can walk and talk. This could be my big ticket to Internet fame and fortune!
2. And then someone will want to marry one, a la William Gibson's novel Idoru. It'll be a real fight when true AGIs are asking for equal rights. But how about before then when someone wants to extend rights to a fancy chatbot with an animation package that we know isn't sentient? Will forming a corporation help or hurt that effort?
We do live in interesting times.
[1] https://www.vogue.com/article/lilmiquela-miquela-sousa-insta...
Software has getting/gotten extremely good at mimicry. Natural motion and facial expressions are in development. I really don't see these things as being that far off into the future where the human is just moving a mouse/hand to find the motion and expressions for a specific sequence. Video game avatars are the best indicator. If you go back 5, 10, 15 years and compare those to what we can do now then extrapolate 5-15 years.
As a customer, I would definitely be better off with an AI Selfie - I'd be happier with my purchases more often and maybe even get some sort of hidden psychological benefit to not looking at unrealistic model bodies. But I'm not sure retailers would stand to benefit much.
What if every time you shopped online you could see a version of yourself you'd indicated you want to be (via a thousand small web interactions) in those clothes?
I think there is a fairly huge middle ground. I wish that REAL models would digitally represent themselves as 3D meshes, so that I could preview digital clothing on them. That would really sell clothes man.
"Subfield" is more correct but it's interesting that a model (and that's not a language model) gets the relation between neural nets, machine learning and AI right, when the majority of the so-called tech press gets it consistenty wrong, e.g. using AI to refer to deep learning in a kind of reverse-synecdoche.
The biggest takeaway for me was that this technology will likely naturally evolve to seeing ourselves in the content and clothing we want. Maybe it's a bit narcissistic to declare publicly, but I have personally seen through my own work the march towards personalization: what's more personal than seeing yourself everywhere doing everything?
Right now there are a lot of folks that are not models tied into this as well: photographers, lighting and set people, makeup, dressers, travel arrangers, fixers, etcs.
I can easily see a near future with the equivalent of Unreal Engine for modling. All sets, lighting, makeup, AND people in picture will be life-like. There will be easily configurable random but realistic auto-posing, etc.
The jobs will all become highly comodified down to low paying jobs for long hours much like the video game industry is today.
And none of the afore mentioned jobs or attendent costs will be required.
As for consumers wanting to "know" the real models and their lives and advantures? Ok well, I'll get off your lawn grandpa. If current trends contue, none of that will matter. Folks already form para-relationships with digital/fantasy people (re: go to any cosplay convention). So the models not being "real" will pose no barrier in the long run.
What remains to be considered is human judgement. But it won't be long. There has been some research on automating the aesthetic placement of the camera within the scene... automatic composition. https://ieeexplore.ieee.org/document/9112197
As a painter, I certainly know there are rules, albeit very 'floppy' ones.
Shudu and her peers _look_ CGI, and I think that's deliberate. Compare the Shudu instagram with the output of a modern GAN. I think this is likely an intentional aesthetic choice. What if part of the appeal of a digital model is it can look both like and not like a person, in a way that is controllable and expressive?
A separate question from the aesthetics is, could it be better to "objectify" something that's already an object, than to do that to a real person?
https://www.instagram.com/p/BnGmYwzF6nR/ https://thispersondoesnotexist.com/
We might be further along towards CGI models than I thought
I do agree that AI models may/can become huge. But there will still be plenty of room for human ones.
It almost feels that if AI models became prevalent the result would be to make some human models even more interesting.
This is just a gut feeling of course.
I can definitely see how a major fast-fashion brand can adopt this practice.
Won't kill Paris and NYC fashion week, but certainly will decrease the number of models that are currently paid for more trivial modelling jobs.
[0] https://kotaku.com/most-pics-in-ikea-catalogues-arent-photos...
We have seen AI (or CGI) being used increasingly in film, music and writing, but the highest forms of these arts are not AI, unless they are AI for AI's sake (i.e. as a novelty).
A fintech firm might now be producing daily stock summaries from AI. A Hollywood studio might make use of CGI in its movies. But the highest art form still makes without AI and will continue so for many decades to come.
There is at least some likelihood that it's precisely the elements or aspects which are not automatable, or which are not automated, which will achive higher status.
edit: *than I can, not "that I can", hah.
So many stories of abuse and mistreatment and them eating tissue paper or being sexualized at 14 keep periodically occurring that eventually enough people will just say forget it and use digital creations.
Beyond the fad and hype sales cycles of fashion perhaps art will flourish again with all of the excess natural beauty that is still in demand. Its still timeless!
Putting real life actors in AAA games has been a thing for years at this point, but now the graphics are so advanced that it will look completely photo real within the next couple console generations. Those game companies put real life actors in their movies because of the audience recognizes them. Same thing will likely happen for fashion as well. If you buy famous models'/celebrities' digital model you can reuse and license that however you want.
1. Hollywood in the Cloud : The progress of computer vision algorithms, game engines like Unreal and massive computation in the cloud mean that in 25 years time you maybe able to produce a Hollywood quality movie by writing code. Unreal engine will render the backgrounds, Deep neural nets will generate the actors voices and faces and code will be used to stitch everything together. This may also include the production of background scores by neural nets primed on music in similar scenes. The number of people needed to produce a film will be cut by 10x and we will see an explosion in film making. Tik Tok is an early example of this.
2. Digital Models: This is connected to the above. You will also see digital models being used on billboards and news readers will be replaced by models like GPT-3 that convert data into narratives and then they are read by digital newsreaders. They may even make it interactive by reading out the most popular tweets or having fake discussion between AI models with different personalities.
3.Lawyers : GPT-3 has given me a lot of confidence in predicting a major disruption to legal research. You can probably semi automate case research and you don't need armies of junior lawyers or para legals to fight cases.
4.Accountants: This relies on the continued improvement in computer vision in the ability to read and interpret printed invoices. More and more transactions will happen via APIs and be shepherded by digital accountants too.
5.Programmers: I am less sure of how programmers will be replaced but there are some obvious avenues. Natural language interfaces could make most front end work obsolete. You don't really need an Uber app if GPT-3 on steroids can understand exactly what you want and then produce a widget on the fly that shows you the appropriate information on demand. Most simple apps will be folded into natural language assistants which means that front end work will go down. What does exist will be designed with the assistance of AI tools. The backend work could also increasingly be subsumed into making a knowledge base that can learn and respond to intelligent queries.
6.Therapists: Smarter NLP models could act as digital therapists. People maybe more comfortable talking to a digital therapist and not be judged by an actual person. They can be given digital bodies and voices to make them more realistic. GPT-3 is way ahead of ELIZA and even in the original ELIZA studies people became quite attached to it. Technology is making people lonely and people may turn to technology to fix it.
7.Fake twitch streamers / Cam models: Synthesis algorithms could become so advanced that some people could become more attractive versions of themselves and create fake model personas that make a lot of money on websites like Twitch.
Our economy and education system are probably unprepared for the scale of disruptions we may see.
I guess this is all dependent on how "in demand" models are...
Usage rights.
Time of models and photographers is cheap - for below supermodel catwalk class it is less than $150/h including the overhead. For catalog/commercial models $50/h including overhead is a good pay.
Buying out of usage rights is expensive. Worldwide buyout for a dozen images for 1 years could easily be $50,000. So instead they get those 12 images for that specific usage type (online) for the time rate + $1
lets make one step further - how about digital actors' images adjusted slightly for any given movie watcher. Can't be done with real people. New tech isn't always "better" (like in "better horse"), it opens/brings in new possibilities/capabilities (like in "car").