Honestly, I expect the number of art roles at gaming companies to drop precipitously over the next couple of years.
Right now you have a creative director who defines the style of art and a bunch of early-to-mid-career artists/designers who then create all the assets you need (and there are a lot of assets, so there's plenty of work to go around). Soon that'll be a creative director defining the style, AI producing most of the assets, and a few designers cleaning up what the AI produces. In a few years, those few designers probably won't be needed.
I think a lot of artists who take commercial art/design type jobs to pay the bills are going to be out of work soon, and that's a serious problem that we should be looking to get ahead of as a society.
On the plus side, this will be a huge boon for indie game devs - they'll be able to create AAA-level assets for their games at extremely low cost. For better or worse (for better and worse, really), the times, they are a-changin'.
Right now the lines are kind of blurred between people who make great art, and people who have the technical skills to create nice graphical imagery.
Lets say that someone gets a character design dropped on their desk, and are told to create some graphics of that character running through a field, climbing a tree, jumping through a window, all using the identical style as the character design sheet. I don't really think there's a lot of real artistry going on there, despite the fact that the person with the skills to perform that work today almost certainly developed those skills while creating their own personal art portfolio.
I have not worked in the gaming industry, but what I hear is that a lot of gaming studios today hire up legions of technically skilled people, and grind them to death in sweatshop-like conditions, while not actually utilizing any of their creative talent.
My hope is that this new tech will empower real artists to massively multiply their productivity, so they can focus on creating new beautiful styles, stories, and personalities for their characters, rather than spending all day painting.
It kind of feels like a Gutenberg Press scenario. Before the Printing Press, you had to have thousands of artists transcribing books, copying the images in great detail, and using their impeccable handwriting to ensure that each page was as legible as their source material. I'm sure that back then it felt like thousands of artists were losing their jobs to these machines, but it ended up creating an explosion of literacy and enabled subsequent generations to publish their work at a scale that was never before possible.
As a non-artist, I'm hesitant to say what's real artistry and what's not, but I definitely agree that there's a more creative component to work of a creative director defining the art style vs. an entry-level artist building assets in that style.
Still, I think one of the things that will be lost is the ability of those entry-level artists to learn and develop so they can get to the point of being qualified for the really creative art jobs. My feeling is that these AI tools will really entrench the existing art-director-level folks in AAA gaming in their roles, and the only path to break into this kind of game art design will be by creating your own games entirely.
As for the sweatshop conditions, I totally agree - the gaming industry is notorious for being awful to its employees, and it's good that that will likely end. On the other hand, is it a plus to end that by just eliminating the jobs entirely? I dunno.
Having working on a bunch of AAA games, having the same number of artists making a lot more assets is a way more likely outcome than having less artists make the same stuff as today - especially if you take in account the trends of making everything have huge worlds with lots of detail.
This is what I was thinking even though I’m totally outside the gaming industry. It seems like the appetite for good quality artwork in video games is absolutely insatiable.
If you can hack out unique stuff in hours, not days, suddenly every building in every city of GTA can have bespoke furniture and stuff that’s absolutely unimaginable to us right now as assigning a dev to spend three days modeling a couch used in one place is absurd.
This is how we get one step closer to actual realistic or extremely detailed environments.
I was playing God of War and the attention to detail in a room filled with treasure, how it reacts when you hit it with your weapon and coins fly everywhere, it was amazing.
Imagine what these same devs could do where their workflow takes 1/10th or 1/100th the time.
I think this is probably true in the short term, but in the long term I think the ability of AI to create art will outpace the need for more new assets. Right now there's definitely cleanup that needs to be done on AI-generated stuff, but the need for that will diminish quickly. I'd be shocked if by the end of the decade we still had artists cleaning up AI work.
Games may need 1000x the assets they do now, but if AI is 10000x faster at creating them than humans are, then you're still going to see a big reduction in the number of humans needed. Eventually we'll get to the point where the game's AI is procedurally generating assets during gameplay (and eventually you'll get AI just generating the games).
>I think a lot of artists who take commercial art/design type jobs to pay the bills are going to be out of work soon, and that's a serious problem that we should be looking to get ahead of as a society.
They can learn this new tools and be 100x more productive and make more money. The AI tool will not replace the artistic taste and creativity , some boring low quality work will be automated but artists can be more productive with this new tools and say finish the artwork for a novel or game much faster, so making more money.
Yeah but if you make artists 100x more productive the natural consequence is that there will be far fewer employed artists.
The senior ones will probably be ok (as you say, they might even get a payrise, given how much impact a single individual will be able to have) but people just joining the industry now are likely screwed.
Say I would like to make a Visual novel but it will take an artist 100 hours to illustrate it, if he can do it 10x faster then I can have him illustrate 10 of my novels with the same money. Nobody loses and my VN readers can enjoy 10x more content.
Btw this AI tools work a lot better if you sketch something first, so I can see a very productive artist doing a rough sketch then using AI then manually touching stuff after.
Also an artist could train the AI with his style and with his characters and keep the model private and use it over and over again.
I think it could mean games have the same art budget but make 100x the art content to compete with studios that use the innovation to slash their headcount..
Automation and better technology massively reduced the fraction of the population that works in farming, whilst simultaneously we ended up with a food surplus. Machine translation definitely had an impact on the human translation industry. Seems likely that this is what will happen for art.
The majority of that work is today already outsourced to art “sweatshops” in cheap cost of living countries. You have one art director telling them to create 100 of this armor and 30 of this creature. The biggest impact will be at the bottom of the ladder
>The majority of that work is today already outsourced to art “sweatshops” in cheap cost of living countries. You have one art director telling them to create 100 of this armor and 30 of this creature. The biggest impact will be at the bottom of the ladder
Do you think that in future the art director will do the work himself using AI? Or he will hire the same people but instead of 100 armors he will ask for 10k armors, more armors more chances to sell them to players.
Seems like unless something changes you’ll still need some people with “taste” going through the hundreds of iterations looking for the right fit. Probably not the director. Still I think the team size will be cut down fairly drastically
I don't think drastically.
Have yu tried the AI tools?
It is not like you can just describe what is in your mind and it will appear, from what I read in people workflow description it is a more involved work, including training a model (I seen someone trained a model for RPG fantasy characters for example), then creating some starting sketch, generating many images, selecting good candidates, then use in-painting to fix out of place stuff, then use photoshop/krita to fix stuff .
I think it is like with programming, we have higher level languages and libraries and the number of developer increased because we can do a lot more stuff then just starting each time from assembly. So in this case the same artists will just produce more, we will not have to wait 6 months for a visual novel to have an update, we will not need to wait years for some indie game to be done , but yeah it is possible that some patt of the money that goes to art will end up for a short period of time in the pockets of soem art director , but competition would fix that.
I'll take you at 1:1 odds that the salary of even the highest paid 10% of art directors in the United States, as reported by the Bureau of Labor Statistics, does not see even a 10 times increase in their salary over 5 years (that is to say, in 2027.m01.08, the BLS reports below 10*$194,130). Loser gives the winner one Snickers bar.
There was an article recently about how AAA games are requiring so much content lately that any development cycles starting today would need to target the PlayStation 6. I think there’s still some room for improvements in productivity.
But is that assuming that content production will continue at its same pace? Unless it was an AI-focused article, it probably wasn't accounting for the changes we're going to see.
There’s no imperatives for selling plastic discs at $79. What that tells is just how much opportunities there are for busywork sharing in console games.
Environment art is another thing that’s already being automated with simpler tools (speedtree etc). No one has time to fill in the trees and rocks for an open world by hand. Games are only getting more demanding as people expect more and more detailed worlds
I personally don't think so, it made incredible progress very fast but it stagnates now. The more precise design you want the less likely you'll be able to generate it. It's usually very nice looking at first but once you look at the details it completely breaks apart, it's a bit like in a dream.
The author seems to be a 3D artist, and claims to have spent 18 hours for this work, that he estimates as 5 days worth, that he thinks not copyrightable. To me it seems to be just raising the bar/granting more productivity for artists with prior skills and abilities to read the room.
My understanding re: copyright was that the raw output of the AI cannot be copyrighted. But his rigging and mesh maps and otherwise creatively-directed work to take that raw material and make it look good? Surely he has copyright on those?
It's just that in theory, anybody else can take the raw output of the AI and do all that extra work on it themselves with no copyright violation.
>So, uh, is every artist going to be out of the job in the next 5 years?
No, probably artists also ask if ChatGPT will replace developers and they can just ask the AI write me the code for a cool RPG game, make it in CoolLang and use NN for NPC AI, also make it super efficient and optimized and similar keywords.
This tools will be used by artists in their workflow, they will be more productive so more high quality art will be created .
Short-sighted employers will probably cut staff. The other more likely scenario is that teams will be asked to produce more work faster by letting AI handle the boilerplate. Creative directors and producers will run a lot more experiments with fresh ideas and then use the expert skills of their teams to produce a cohesive, polished product.
The end result may be that we get a ton of mediocre, high-fructose-corn-syrup art at first. The hype for that will last about three months. We’ll also see an increase in typical fare from big studios as they figure out how to monetize long-tail content. And then occasionally there will be a true work of art or a new form of art we haven’t seen yet.
