Except, right now, they absolutely are. The tools are largely as you describe - though thinking about it, I think I’d describe it more as an airbrush vs a paint brush - but that’s not the way that upper management sees it for the most part, and not how the average supporter of GenAI sees it even if they don’t recognize that that’s their view. Both of these groups see it as a way to cut costs by reducing manpower, even if the GenAI folk don’t recognize that that’s what their stance is (or refuse to accept it). It’s the same as in the programming side of the conversation: vibe coders and prompt generators being hired instead of skilled professionals who can actually use the tools where they’re truly useful. Why pay an artist or programmer to do the work when I can just ask an LLM trained on stolen work to do it for me instead.
I read a great post probably a year ago now from somebody who works for a movie studio on why the company has banned hiring prompters. The short of it is, they hired on a number of prompters to replace some jobs that would normally be filled by artists as a test to see if they could reduce their staff while maintaining the same levels of production. What they found was that prompters could produce a massive volume of work very quickly. You ask the team for pictures of a forest scene and the artists would come back in a week with a dozen concepts each while the prompters had 50 the next day. But, if you asked them to take one of their concept pieces and do something like remove the house in it or add people in the foreground, they’d come back the next day with 50 new concept pieces but not the original. They couldn’t grasp the concept of editing and refining an image, only using GenAI to generate more with a new set of prompt parameters, and therefore were incapable of doing the work needed that an artist could do.
A feel-good story for artists showing what AI is actually capable of and what it isn’t, except for one thing: the company still replaced artists with AI before they learned their lesson, and that’s the phase most of the world is in right now and will probably continue to be in until the bubble bursts. And as Alanah Pierce so eloquently put it when talking about the record setting year over year layoffs in the gaming industry (each year has been worse than during the 2008 financial crash): “Most of those people will never work in games again. There’s just too many people out of work and not enough jobs to go around.” These companies currently in the fuck around phase will find out eventually, but by then it won’t matter for many people. They’ll never find a job in their field in time and be forced into other work. Art is already one of the lowest paying jobs for the amount of effort and experience required. Many artists who work on commissions do so for less than minimum wage, and starting wages in the game industry for artists haven’t increased since I was looking at jobs in the field 15 years ago.
More that an existing smaller studio doesn’t have to sell their soul to a publisher (or get lucky) to survive. They can more safely make a “big” game without going AAA.
My observation is that there’s a “sweet spot” for developers somewhere around the Satisfactory (Coffee Stain) size, with E33 at the upper end of that, but that limits their audience and scope. If they can cut expensive mocap rigs, a bunch of outsourced bulk art, stuff like that with specific automation, so long as they don’t tether themselves to Big Tech AI, that takes away the advantage AAAs have over them.
A few computer generated textures is the first tiny step in that direction.
So no. AI is shit at replacing artists. Especially in E33 tier games. But it’s not a bad tool to add to their bucket, so they can do more.
Right, so the barrier was that they had to pay for this “outsourced bulk art”, and now with AI they don’t have to. It looks like we are in agreement when I say “I’m glad AI is democratizing the ability for the little guys like you and me to not pay artists for art”?
It takes less time for the actual in house artists to use GenAI with a dataset trained with the company’s own style to generate “bulk art” than it takes them to manage an outsourced company doing the same thing.
Sauce: work in gaming, just talked about this with our art producer.
The outsourcing work is literally “make this texture we made ourselves by hand look like it was snowing” type of shit. You can use GenAI and have it done in 30 minutes or spend 2 hours talking back and forth with the outsourcing partner in 10 minute intervals over a week - interrupting your flow every time.
I think AI is too dumb, and will always be too dumb, to replace good artists.
I think most game studios can’t afford full time art house across like 30 countries, nor should they want the kind of development abomination Ubisoft has set up. That’s what I’m referring to when I say “outsourced”; development that has just gotten too big, with too many people and too generic a target market. And yes, too many artists working on one game.
I think game artists should have a more intimate relationship with their studio, like they did with E33.
And it’d be nice for them have tools to make more art than they do now, so they can make bigger, richer games, quicker, with less stress and less financial risk. And no enshittification that happens when their studio gets too big.
The implication here is that you can gain manpower without hiring more men, no?
One builder only uses hand tools, other uses power tools.
That’s the difference, nobody is hiring less people because the tools are better.
Except, right now, they absolutely are. The tools are largely as you describe - though thinking about it, I think I’d describe it more as an airbrush vs a paint brush - but that’s not the way that upper management sees it for the most part, and not how the average supporter of GenAI sees it even if they don’t recognize that that’s their view. Both of these groups see it as a way to cut costs by reducing manpower, even if the GenAI folk don’t recognize that that’s what their stance is (or refuse to accept it). It’s the same as in the programming side of the conversation: vibe coders and prompt generators being hired instead of skilled professionals who can actually use the tools where they’re truly useful. Why pay an artist or programmer to do the work when I can just ask an LLM trained on stolen work to do it for me instead.
I read a great post probably a year ago now from somebody who works for a movie studio on why the company has banned hiring prompters. The short of it is, they hired on a number of prompters to replace some jobs that would normally be filled by artists as a test to see if they could reduce their staff while maintaining the same levels of production. What they found was that prompters could produce a massive volume of work very quickly. You ask the team for pictures of a forest scene and the artists would come back in a week with a dozen concepts each while the prompters had 50 the next day. But, if you asked them to take one of their concept pieces and do something like remove the house in it or add people in the foreground, they’d come back the next day with 50 new concept pieces but not the original. They couldn’t grasp the concept of editing and refining an image, only using GenAI to generate more with a new set of prompt parameters, and therefore were incapable of doing the work needed that an artist could do.
A feel-good story for artists showing what AI is actually capable of and what it isn’t, except for one thing: the company still replaced artists with AI before they learned their lesson, and that’s the phase most of the world is in right now and will probably continue to be in until the bubble bursts. And as Alanah Pierce so eloquently put it when talking about the record setting year over year layoffs in the gaming industry (each year has been worse than during the 2008 financial crash): “Most of those people will never work in games again. There’s just too many people out of work and not enough jobs to go around.” These companies currently in the fuck around phase will find out eventually, but by then it won’t matter for many people. They’ll never find a job in their field in time and be forced into other work. Art is already one of the lowest paying jobs for the amount of effort and experience required. Many artists who work on commissions do so for less than minimum wage, and starting wages in the game industry for artists haven’t increased since I was looking at jobs in the field 15 years ago.
“Replace tools, not jobs” is the best way to use AI.
And also the one that works the best both for people and businesses.
Replacing jobs feels cooler for bosses though…
More that an existing smaller studio doesn’t have to sell their soul to a publisher (or get lucky) to survive. They can more safely make a “big” game without going AAA.
My observation is that there’s a “sweet spot” for developers somewhere around the Satisfactory (Coffee Stain) size, with E33 at the upper end of that, but that limits their audience and scope. If they can cut expensive mocap rigs, a bunch of outsourced bulk art, stuff like that with specific automation, so long as they don’t tether themselves to Big Tech AI, that takes away the advantage AAAs have over them.
A few computer generated textures is the first tiny step in that direction.
So no. AI is shit at replacing artists. Especially in E33 tier games. But it’s not a bad tool to add to their bucket, so they can do more.
Right, so the barrier was that they had to pay for this “outsourced bulk art”, and now with AI they don’t have to. It looks like we are in agreement when I say “I’m glad AI is democratizing the ability for the little guys like you and me to not pay artists for art”?
It takes less time for the actual in house artists to use GenAI with a dataset trained with the company’s own style to generate “bulk art” than it takes them to manage an outsourced company doing the same thing.
Sauce: work in gaming, just talked about this with our art producer.
The outsourcing work is literally “make this texture we made ourselves by hand look like it was snowing” type of shit. You can use GenAI and have it done in 30 minutes or spend 2 hours talking back and forth with the outsourcing partner in 10 minute intervals over a week - interrupting your flow every time.
I think AI is too dumb, and will always be too dumb, to replace good artists.
I think most game studios can’t afford full time art house across like 30 countries, nor should they want the kind of development abomination Ubisoft has set up. That’s what I’m referring to when I say “outsourced”; development that has just gotten too big, with too many people and too generic a target market. And yes, too many artists working on one game.
I think game artists should have a more intimate relationship with their studio, like they did with E33.
And it’d be nice for them have tools to make more art than they do now, so they can make bigger, richer games, quicker, with less stress and less financial risk. And no enshittification that happens when their studio gets too big.