A typical point that I severely miss from most discussions about AI is what it means for future artists or, in this case, future actors. And therefore what it means for us as a society.
By taking the art from the artists, regardless of whether it’s an actor, illustrator, author, etc…, the way it is done currently, we will see much fewer people who will even try to learn these skills, or share them. At some point there won’t be anything new anymore.
but isn’t the artistic field already a lottery when it comes to making a living doing it? Maybe I have the wrong impression, but I feel like if “I very likely won’t be able to make a living doing this” actually discouraged new art from getting created, it already would be doing that.
Only if you’re looking at the very top of the profession, like people who hit it big as stars. There are a lot of other levels of employment and success short of Banksy or Beeple level.
My hope is that deep-faking tech might actually help lower levels of the profession, even if it’s at the expense of those at the top who get huge amounts of money because of how famous their face is.
Imho, Studios don’t even need to copy a famous actor’s face… just create a face of a person who doesn’t exist and make it into a new famous character by stamping it into a good (even if not top famous) actor.
That’s true, it’s entirely uneccessary for people like Tom cruise to exist.
So far it looks like that’s not their plan, though, with the offer to digitize extras for a one-time payment of $200. So they’ll just entirely replace extras forever with AI for what they’d normally make for 2-3 days of shooting.
That’s fair; the person I first responded to seemed to be discussing the “fine art” part.
How worried are you personally about these more advanced machine learning tools? Just a month or so ago I was playing a Pathfinder (rpg) video game and didn’t care or any of the built-in avatar images, so I hopped onto one of the websites that make an image based on a text prompt to make me an image that matched my character and it took a few times to get the right wording but in the end I got a pretty good image out of it. I vaguely know how it works. (vaguely is doing a lot of heavy lifting in that sentence) and it still seemed kind of like magic.
the clients (no offense) tend to be less professional
I don’t know what you mean by this precisely, but the “pretty good” end result I mentioned had a hand that melted into the sword-- so if you meant “low standard” then yeah, guilty as charged, haha. However, more interesting to me is that I would have never in 1000 years have paid someone to do that for me-- I just would have been low-level annoyed that my character and the avatar looked different the entire game.
I find the “they didn’t have permission to train from” argument is complete bunk. That’s not a right granted by intellectual property laws; there is no “right to control who learns from a work”.
What needs to happen is society (especially US society) needs to stop linking “working” and “enjoying a comfortable life”. Technology is coming for all our jobs, and the sooner we accept that and prepare for it, the better we’ll be when it happens.
I find the “they didn’t have permission to train from” argument is complete bunk. That’s not a right granted by intellectual property laws; there is no “right to control who learns from a work”.
Yeah, but is an AI LEGALLY learning? Or is it just a machine that spits out output based on its inputs? In that case, use of the work as input isn’t allowed under the copyright, which is that the work be used by reading it.
All these comparisons between what an AI is doing and what a human does when reading/learning/etc are not a given in a court of law. We don’t have any rulings yet that an AI is actually “learning” like a human when it is “trained.”
“Training” an AI is building a tool. A tool that can be used to profit. Can artistic works be used to build a for-profit tool without permission?
This is something that needs to be decided, and it will be decided in a way that whatever the rules are for AI can’t be applied to a human. Meaning if there is a requirement for permission for use in machine learning, that won’t change that a human can learn from it. So the comparison is pointless, because there is no way the courts are going to rule that these things are legally indistinguishable from people.
In the meantime, back to the original, there ARE precedents for use of performance because of recordings. That’s why the studios wanted that in the contract, they KNOW they cannot manipulate a person’s performance through AI without their express written permission. Is it REALLY so hard to believe this can be applied to writing or art? That they can’t use writing or art without the artist’s express permission.
We may see a new kind of copyright soon that specifically disallows use for AI, and another that is open for use with AI. Something to replace Creative Commons on the internet.
By that I mean that some dude looking for a game avatar or whatever isn’t as likely to be someone used to contracting people to do professional creative work for them. Professional clients who are accustomed to hiring creatives to do work for them are more likely to:
be quick to provide feedback and respond to emails
have feedback that is clear and actionable
communicate professionally
pay on time
be willing to pay a down payment and sign a contract
comprehend the hard work that goes into this stuff and value my time accordingly
don’t try to push a project (either intentionally or otherwise) into what is known as “scope creep” wherein they jockey for additional work outside of the initially agreed upon scope of work
And lots of other little things like that.
Am I saying that you are like that? No. But having done creative work for north of 15 years this is my informed opinion based on a lot of experience in this field.
I find the “they didn’t have permission to train from” argument is complete bunk. That’s not a right granted by intellectual property laws; there is no “right to control who learns from a work”.
That’s your opinion and you might feel differently if you had spent years working hard to achieve something in this specific field.
What needs to happen is society (especially US society) needs to stop linking “working” and “enjoying a comfortable life”. Technology is coming for all our jobs, and the sooner we accept that and prepare for it, the better we’ll be when it happens.
This I fully agree with. And I wouldn’t even necessarily have a problem with AI destroying creative jobs if it meant I was now more free to pursue a life of spending time doing things that I was passionate about because some kind of UBI or whatever was making that possible for me and others.
Like I kinda mentioned earlier, I don’t think society is in a good place to fight for this on at least the short term. Basically not until things get really bad. What I expect will happen for now is most all of the windfalls from automation will be siphoned up to to the upper class and corporations and wages will continue to stagnate for the working class and income inequality will continue to skyrocket.
No. Before you could actually live (albeit barely) on being a designer or an illustrator, small gig actor or author for articles, musician for jingles, etc… Even when you weren’t the best and famous already. Artists are already seeing this slipping away and with further advances in AI you really do need to be one of the already famous people to do these types of art as a viable job.
That’s the problem - it take a lot of practice and experience to get really good at graphic design or illustration. When people are paying you to do it, you can afford to do it all day. If not, you need to spend the majority of your time doing something else, so it takes longer to advance in skill. I see this in my own field with hobbyists/people who do art on the side vs people who do it full time.
It will mean that not only do you need to compete with your peers, you’ll need to compete forever with all the best talent that has ever worked.
And those talents, at a certain point, will cost less. They’ll be able to do more for less money because they’ll be on to other things or dead, and thus are handling their living (or not) expenses differently. While you’ll still need an apartment near the studios and food to survive.
There’s no real up side for 99.99% of people. The only ones who will make any real money from these changes are the executives and producers.
Most artists can’t earn their entire livelihood by their craft alone. Even those considered good, in most cases, need a main job.
But even the little money you make from your art can at least pay for art supplies (which are very expensive). Learning to be a good in your craft costs an enormous amount of patience, time and money as well. With no money at all to be made out of it, no commissions, and your work immediately flowing into the AI pipeline, new artists will be further discouraged from even trying to hone that craft.
You may very well be right on the money here, but I find it at least plausible that a market for “human-made” art becomes a thing if computer-made art becomes a thing.
It will only be rich people who can afford to do that, then. It won’t be a job anymore and even less likely to be a profitable endeavour for the many who can’t just pour all their time and money into a hobby just to become that good at it one day.
That’s not necessarily true. Certainly plausible, but just as plausible as it working out like “cage free” eggs, where a perceived value pushes the market into a direction that it wouldn’t go for purely financial reasons.
These are my thoughts exactly as well. I find it especially infuriating how in some discussions and articles about AI people try to spin a tale of artists being these allegedly elitist “bourgeois” individuals. And with AI now supposedly “the little man” can finally unfold their creative potential. In truth I suspect it’s more the other way around.
It looks to me like the only kinds of people who could afford to get into my line of work in the future are people with rich parents or something
Someone from a not so well-off background but with dedication and grit was more likely to get their feet into illustration, für example, then they are now.
Instead people who already probably come from a privileged background (PC, technical knowledge, money to pay for AI credits) can just swarm the market without needing to dedicate much time at all.
I saw a comment on reddit a while back on the topic that I thought really hit the nail on the head in terms of the typical discourse I keep seeing on AI with respect to art and other creative work
“If they [tech bros] don’t value art that’s fine, but it’s sad that people who don’t value art are the ones who think they should be deciding what direction art goes in.”
A typical point that I severely miss from most discussions about AI is what it means for future artists or, in this case, future actors. And therefore what it means for us as a society.
By taking the art from the artists, regardless of whether it’s an actor, illustrator, author, etc…, the way it is done currently, we will see much fewer people who will even try to learn these skills, or share them. At some point there won’t be anything new anymore.
Maybe I’m overlooking something, but isn’t the actual change that doing these things will no longer be a viable way to earn a living?
deleted by creator
but isn’t the artistic field already a lottery when it comes to making a living doing it? Maybe I have the wrong impression, but I feel like if “I very likely won’t be able to make a living doing this” actually discouraged new art from getting created, it already would be doing that.
Only if you’re looking at the very top of the profession, like people who hit it big as stars. There are a lot of other levels of employment and success short of Banksy or Beeple level.
My hope is that deep-faking tech might actually help lower levels of the profession, even if it’s at the expense of those at the top who get huge amounts of money because of how famous their face is.
Imho, Studios don’t even need to copy a famous actor’s face… just create a face of a person who doesn’t exist and make it into a new famous character by stamping it into a good (even if not top famous) actor.
That’s true, it’s entirely uneccessary for people like Tom cruise to exist.
So far it looks like that’s not their plan, though, with the offer to digitize extras for a one-time payment of $200. So they’ll just entirely replace extras forever with AI for what they’d normally make for 2-3 days of shooting.
deleted by creator
That’s fair; the person I first responded to seemed to be discussing the “fine art” part.
How worried are you personally about these more advanced machine learning tools? Just a month or so ago I was playing a Pathfinder (rpg) video game and didn’t care or any of the built-in avatar images, so I hopped onto one of the websites that make an image based on a text prompt to make me an image that matched my character and it took a few times to get the right wording but in the end I got a pretty good image out of it. I vaguely know how it works. (vaguely is doing a lot of heavy lifting in that sentence) and it still seemed kind of like magic.
deleted by creator
I don’t know what you mean by this precisely, but the “pretty good” end result I mentioned had a hand that melted into the sword-- so if you meant “low standard” then yeah, guilty as charged, haha. However, more interesting to me is that I would have never in 1000 years have paid someone to do that for me-- I just would have been low-level annoyed that my character and the avatar looked different the entire game.
I find the “they didn’t have permission to train from” argument is complete bunk. That’s not a right granted by intellectual property laws; there is no “right to control who learns from a work”.
What needs to happen is society (especially US society) needs to stop linking “working” and “enjoying a comfortable life”. Technology is coming for all our jobs, and the sooner we accept that and prepare for it, the better we’ll be when it happens.
Yeah, but is an AI LEGALLY learning? Or is it just a machine that spits out output based on its inputs? In that case, use of the work as input isn’t allowed under the copyright, which is that the work be used by reading it.
All these comparisons between what an AI is doing and what a human does when reading/learning/etc are not a given in a court of law. We don’t have any rulings yet that an AI is actually “learning” like a human when it is “trained.”
“Training” an AI is building a tool. A tool that can be used to profit. Can artistic works be used to build a for-profit tool without permission?
This is something that needs to be decided, and it will be decided in a way that whatever the rules are for AI can’t be applied to a human. Meaning if there is a requirement for permission for use in machine learning, that won’t change that a human can learn from it. So the comparison is pointless, because there is no way the courts are going to rule that these things are legally indistinguishable from people.
In the meantime, back to the original, there ARE precedents for use of performance because of recordings. That’s why the studios wanted that in the contract, they KNOW they cannot manipulate a person’s performance through AI without their express written permission. Is it REALLY so hard to believe this can be applied to writing or art? That they can’t use writing or art without the artist’s express permission.
We may see a new kind of copyright soon that specifically disallows use for AI, and another that is open for use with AI. Something to replace Creative Commons on the internet.
By that I mean that some dude looking for a game avatar or whatever isn’t as likely to be someone used to contracting people to do professional creative work for them. Professional clients who are accustomed to hiring creatives to do work for them are more likely to:
And lots of other little things like that.
Am I saying that you are like that? No. But having done creative work for north of 15 years this is my informed opinion based on a lot of experience in this field.
That’s your opinion and you might feel differently if you had spent years working hard to achieve something in this specific field.
This I fully agree with. And I wouldn’t even necessarily have a problem with AI destroying creative jobs if it meant I was now more free to pursue a life of spending time doing things that I was passionate about because some kind of UBI or whatever was making that possible for me and others.
Like I kinda mentioned earlier, I don’t think society is in a good place to fight for this on at least the short term. Basically not until things get really bad. What I expect will happen for now is most all of the windfalls from automation will be siphoned up to to the upper class and corporations and wages will continue to stagnate for the working class and income inequality will continue to skyrocket.
No. Before you could actually live (albeit barely) on being a designer or an illustrator, small gig actor or author for articles, musician for jingles, etc… Even when you weren’t the best and famous already. Artists are already seeing this slipping away and with further advances in AI you really do need to be one of the already famous people to do these types of art as a viable job.
That’s the problem - it take a lot of practice and experience to get really good at graphic design or illustration. When people are paying you to do it, you can afford to do it all day. If not, you need to spend the majority of your time doing something else, so it takes longer to advance in skill. I see this in my own field with hobbyists/people who do art on the side vs people who do it full time.
It will mean that not only do you need to compete with your peers, you’ll need to compete forever with all the best talent that has ever worked.
And those talents, at a certain point, will cost less. They’ll be able to do more for less money because they’ll be on to other things or dead, and thus are handling their living (or not) expenses differently. While you’ll still need an apartment near the studios and food to survive.
There’s no real up side for 99.99% of people. The only ones who will make any real money from these changes are the executives and producers.
I do not disagree with anything you said.
Most artists can’t earn their entire livelihood by their craft alone. Even those considered good, in most cases, need a main job.
But even the little money you make from your art can at least pay for art supplies (which are very expensive). Learning to be a good in your craft costs an enormous amount of patience, time and money as well. With no money at all to be made out of it, no commissions, and your work immediately flowing into the AI pipeline, new artists will be further discouraged from even trying to hone that craft.
You may very well be right on the money here, but I find it at least plausible that a market for “human-made” art becomes a thing if computer-made art becomes a thing.
It will only be rich people who can afford to do that, then. It won’t be a job anymore and even less likely to be a profitable endeavour for the many who can’t just pour all their time and money into a hobby just to become that good at it one day.
That’s not necessarily true. Certainly plausible, but just as plausible as it working out like “cage free” eggs, where a perceived value pushes the market into a direction that it wouldn’t go for purely financial reasons.
deleted by creator
These are my thoughts exactly as well. I find it especially infuriating how in some discussions and articles about AI people try to spin a tale of artists being these allegedly elitist “bourgeois” individuals. And with AI now supposedly “the little man” can finally unfold their creative potential. In truth I suspect it’s more the other way around.
Someone from a not so well-off background but with dedication and grit was more likely to get their feet into illustration, für example, then they are now.
Instead people who already probably come from a privileged background (PC, technical knowledge, money to pay for AI credits) can just swarm the market without needing to dedicate much time at all.
I saw a comment on reddit a while back on the topic that I thought really hit the nail on the head in terms of the typical discourse I keep seeing on AI with respect to art and other creative work