I guess, but thats kind of misleading. AI literally studying and storing the data of billions of pieces of art to find patterns is not equivalent to a human creating art after a lifetime of seeing or being inspired by art.
AI literally studying and storing the data of billions of pieces of art to find patterns is not equivalent to a human creating art after a lifetime of seeing or being inspired by art.
But how is it different? The “brain” of the AI doesn’t really contain the art it was trained on. Instead it has simply learned some patterns that are common in some source material.
What if, instead of using AI to create lookalikes of known art, emergent behavior was exploited to create art, where you could not deduce what the inspiration was? For example completely new styles of art that yet do not exist? Would that still be plagiarizing?
Where is the line that defines that some content is original and another is not?
I think the line is drawn when a machine is doing the processing. I disagree that the brain of the AI doesn’t really contain the art, as even though this might be literally true, there have been instances of AI generating images that are extremely similar to source images from its dataset (given the right prompt).
I think that any art AI creates is a form of plagiarism, as the only reason it can create anything at all is because it has deduced patterns from its dataset of existing art. Without that other art, it literally could not create anything. In regards to an AI creating a new style of art, I don’t think that will ever happen without sentient intelligence. AI doesn’t have creativity, anything it creates is a combination of data on other art pieces. Also, I think that if the AI’s deduction was observed and compared with it’s prompt, you absolutely could find out exactly what the inspiration was for its new art.
Now if an AI achieved true sentient intelligence, perhaps at a certain scale it would overcome these issues and be able to create original art.
I think the human brain is a mysterious art generator that is much more complex than a large language model finding patterns.
The difference between AI art and human art is that humans can do art by picking up a pencil, maybe even some coloured ones, and a paper. I can ask a kid to draw a dog or something, and they’ll draw some thing that kind of looks like a dog (or not at all lol). They wouldn’t really need a drawing of a dog (or not even a picture sometimes) to draw it.
An artist could also look at a dog and decide to draw it, whilst adding other features that aren’t necessarilly on the dog they’re drawing to give it a style.
An AI artist cannot look at something and “draw” it. We could make something that can take pictures, but that wouldn’t really be drawing. Usually, if someone would make a program to “draw” something, that program would also have to follow the instructions of an artist in order to draw. It also can’t decide on it’s own what it needs to draw, it doesn’t get inspired.
Humans do not need to have other people’s drawings in order to draw. There has to have been a first person who decided to draw, then others followed suite and developped techniques until art has evolved into what we know it is today
AI does need other people’s drawings. It needs it in order to replicate a certain style to make an image a drawing and not just a picture. It can also “create” a new style or technique, but it would never be able to draw unless it analysed pictures beforehand.
Maybe something similar to how humans learn is if it was an AI that would do random scribbles, and if it ressembles whatever it’s trying to draw, it would get positive feedback? That would take a lot of time though, so that’s not what most developpers went for, is it?
I do believe AI art can be used as a good tool, but it’s clear that the goal so far has been to try to find a way to replace the place that artists have in society currently. Kinda sad about it tbh
But yeah AI art and humans do not make art in the same way. Humans don’t need another drawing to draw (at best a reference). They can get inspired by other drawings, but AI can’t. It learns “what” is a drawing through a bunch of pictures, then tries to show something that looks like a drawing. It doesn’t actually draw.
I guess, but thats kind of misleading. AI literally studying and storing the data of billions of pieces of art to find patterns is not equivalent to a human creating art after a lifetime of seeing or being inspired by art.
But how is it different? The “brain” of the AI doesn’t really contain the art it was trained on. Instead it has simply learned some patterns that are common in some source material.
What if, instead of using AI to create lookalikes of known art, emergent behavior was exploited to create art, where you could not deduce what the inspiration was? For example completely new styles of art that yet do not exist? Would that still be plagiarizing?
Where is the line that defines that some content is original and another is not?
I think the line is drawn when a machine is doing the processing. I disagree that the brain of the AI doesn’t really contain the art, as even though this might be literally true, there have been instances of AI generating images that are extremely similar to source images from its dataset (given the right prompt).
I think that any art AI creates is a form of plagiarism, as the only reason it can create anything at all is because it has deduced patterns from its dataset of existing art. Without that other art, it literally could not create anything. In regards to an AI creating a new style of art, I don’t think that will ever happen without sentient intelligence. AI doesn’t have creativity, anything it creates is a combination of data on other art pieces. Also, I think that if the AI’s deduction was observed and compared with it’s prompt, you absolutely could find out exactly what the inspiration was for its new art.
Now if an AI achieved true sentient intelligence, perhaps at a certain scale it would overcome these issues and be able to create original art. I think the human brain is a mysterious art generator that is much more complex than a large language model finding patterns.
The difference between AI art and human art is that humans can do art by picking up a pencil, maybe even some coloured ones, and a paper. I can ask a kid to draw a dog or something, and they’ll draw some thing that kind of looks like a dog (or not at all lol). They wouldn’t really need a drawing of a dog (or not even a picture sometimes) to draw it.
An artist could also look at a dog and decide to draw it, whilst adding other features that aren’t necessarilly on the dog they’re drawing to give it a style.
An AI artist cannot look at something and “draw” it. We could make something that can take pictures, but that wouldn’t really be drawing. Usually, if someone would make a program to “draw” something, that program would also have to follow the instructions of an artist in order to draw. It also can’t decide on it’s own what it needs to draw, it doesn’t get inspired.
Humans do not need to have other people’s drawings in order to draw. There has to have been a first person who decided to draw, then others followed suite and developped techniques until art has evolved into what we know it is today
AI does need other people’s drawings. It needs it in order to replicate a certain style to make an image a drawing and not just a picture. It can also “create” a new style or technique, but it would never be able to draw unless it analysed pictures beforehand.
Maybe something similar to how humans learn is if it was an AI that would do random scribbles, and if it ressembles whatever it’s trying to draw, it would get positive feedback? That would take a lot of time though, so that’s not what most developpers went for, is it?
I do believe AI art can be used as a good tool, but it’s clear that the goal so far has been to try to find a way to replace the place that artists have in society currently. Kinda sad about it tbh
But yeah AI art and humans do not make art in the same way. Humans don’t need another drawing to draw (at best a reference). They can get inspired by other drawings, but AI can’t. It learns “what” is a drawing through a bunch of pictures, then tries to show something that looks like a drawing. It doesn’t actually draw.