A.I. Art Is Destroying My Life Purpose

[AtheisticNonduality]

Yeah, you wanna see the "Atheistic bear" and the "fire owls" in his journal?

[ZzzleepingBear]

11 hours ago, Space said:

Yes this is an important point. AI artists are good at generating artwork but they lack the ability to add subtle and nuanced details which are often an important part of the composition. Maybe its just the eyes of a face looking just slightly in the wrong direction or like you said a small extra details in the background.

However, in a lot of circumstances, I would say that these subtle nuances are not really that important for the client. Unless the detail make or breaks the image, a client will be happy with the result. Most professional clients are not looking for extremely specific images. Mostly what happens is a client will give the artist a fairly simple brief and its up to the artist to come up with the composition, details and make all the decisions. 

But in the not too distant future, if an editorial client has 50 artwork options to choose from based on the title of their new article, they're not going to care about some small detail in image number 18. They'll just choose a different piece.

Yeah I agree with most of what you mention here. This AI tool will definitely have some impact on the way we view art, and the direction some styles might develop. I stil see alot of limitation and differences to who and how someone may use this tool. For example, those who have little to no artistic understanding or experience may be thrilled in creating tons of cool images in general, but will lack the knowledge in how to cleanup their images and be stuck with possible imperfections. While proffesional artists would have to compromise on their own abilities and artistic expression to spend time to cleanup their images, so it may suck the fun out of their proffesion if they deside to speed up their workflow by searching for a satisfying randomized image to work on.

Realistically imo where this AI are now, some art-styles may be in danger, while most don't. And for those styles who are in danger to be overly used,  may turn into a new meta direction where the AI generated images may struggle more to replicate what is consider to be interesting apart from what has become over popularized.

The compromising that has to be accepted when using this AI in general, may be too frustrating for established artists who have a strong will to take their art in a certain direction and who want to be the one in control of the final product, rather than relying on trying to gain better control over a AI tool that are in constant development in how it may randomize it's image of choice.

There are alot of variables to consider or put more or less emphasis on when it comes to generate a certain feeling when creating art. So if all those choices are in the hand of the AI when creating a new image, there will always be a disatisfaction attached to a portion of the generated image no matter how many times you reroll the dice of the rendering  (depending on your experience you have as an artist). I can possibly see some artists who may dislike the project they are working on, to have less care of the final product, and then have as their main goal set to get it all done and over with as fast as possible. Or if an established publisher decides to settle with more of the AI generated images for cheaper work costs, there will eventually be a noticeably difference in the end product that will be recognized overtime as it spreads.  

Anyways, I just ramble on here, but it's a interesting topic with no clear definitive answears. I guess time will tell, but I think there is plenty of room to be optimistic regardless of how fast this development seem to go.

Edited September 1, 2022 by ZzzleepingBear

[zurew]

AI is most likely gonna replace most artists in the next decade or so, because the current quality is just the start, now more and more company see a big business in it, and the competition is getting bigger and stronger and it drives the development pretty well.

 

I haven't seen any good argument yet, why an AI wouldn't be able to do 99% or 100% of the things in the future, that human artists do right now. What structural limitation an AI has that won't allow it to do certain things that a human artist can? Before answering these questions we have to remember some things here:

• Depending on what company we are talking about, you can very very specifically give information to the AI what you want to create (the same way if you would go to an artist) 
• Given enough training time and data, it can basically learn any style 
• It can generate photorealistic images right now
• When you generate an image you can further edit that image with the AI, you can give specific instructions what you want to change and how you want to modify it or if you want to create new images that are very similar to the image that you created before, you can do that as well
• Most people will be able to use this AI, without any specific qualification or knowledge needed to create art, and people will use it, because it will be much cheaper than to hire an artist

 

12 hours ago, AtheisticNonduality said:

A.I. "art" is pretty bad, even from the advanced programs. So make your work exceptional.

There are some bad images, but there are some beautiful images as well, so structurally its capable to generate stunning images if it is trained on enough data. Also, this is just the beginning, as time goes by it will get better and better and more advanced.

10 hours ago, AtheisticNonduality said:

Yeah, you wanna see the "Atheistic bear" and the "fire owls" in his journal?

What about these images?

 

Edited September 1, 2022 by zurew

[Leo Gura]

Look, dude, artists are never gonna be replaced.

Just use the AI to help you make even better art.

This is not bad news, this great news! Humans and AI will cooperate to create amazing stuff.

You're looking at this whole thing with such a narrow perspective. There is a nearly endless need that humans have for art. Your job is not so much to put pencil to paper but to come up with art projects that are really worthwhile. Serious art projects require vision, taste, and leadership. AI will not replace that.

You have a rare opportunity here to be on the cutting edge of art, if you have some vision and play your cards right. You could makes millions and even billions here.

Imagine being able to make a graphic novel using AI by giving it the right prompts.

Edited September 1, 2022 by Leo Gura

[Extreme Z7]

1 hour ago, Leo Gura said:

Serious art projects require vision, taste, and leadership. AI will not replace that.

Couldn't have put it better myself.

Keep in mind that these AI bots have no idea of what the actual cultural landscape of humanity looks like which is something actual visionary creators take into account when thinking up new projects. AI only has text prompts to start from and an admittedly-complex system of probabilistic drives that lend them to use certain images over others. But it doesn't really know anything about the actual psychological impact that specific AI creative decisions have on the audience looking at its work. (In short, it only naively produces a result but doesn't really care about its meaning or value)

Some AI art may look hyper-realistic and "professional", for sure. But if my life-long interest in video gaming has taught me anything, it's that graphics isn't everything.
 

[Danioover9000]

@Space

On 8/31/2022 at 8:25 AM, Space said:

Over the past few weeks and months I’ve been getting increasingly concerned and admittedly very depressed about the onset of A.I art and its impact on professional artists like myself.

In 2021 I escaped wage slavery, quit my 9-5, and became a professional freelance editorial illustrator. But only after literally years of hard work and practice. I’ve put thousands of hours into this. Endless nights of grind to get where I’m at now. Only to now recognise that in the not too distant future my work could easily be replaced by an A.I art generator.

Some of you might think, ‘surely A.I isn’t that good yet?’. I’m telling you it is. It is truly exceptional. And that’s coming from myself as an artist with a critical eye who knows what is good and isn’t. And it’s only going to improve. It’s definitely stronger in some areas vs others, but on the whole it’s exceptionally good already. In 2,3 or 5 years how good will it be? And that’s my issue. Sure, maybe it’s not taking my jobs right now, but in a few years time it very likely will be. 

And that’s the depressing part. Whats the point in continuing on in my line of work (which is already very competitive) if the majority of it will be taken up by a.i generators. Very depressing. I’m struggling to get myself out of this state. Feeling very down and have no motivation to work. My wonderful vision for my life has basically crumbled in front of my eyes. I genuinely feel purposeless which is a dangerous state for a man to be in. The only thing I’ve ever been skilled/talented at (art and drawing) is now slowly becoming useless. 

And look, I’m exposed to this stuff every day. I know what’s out there. I know what kind of art is being made by A.I. And if there’s anyone who would deny the competency of A.I art generators it would be me! It’s in my agenda to deny their capabilities and talk badly about them. But from what I’m seeing, they will be taking a significant large amount of art jobs in the coming years.

So there are obviously areas of art which A.I artists will easily take over and there are domains which it won’t. 

Stock images and photography - this will be the first to go. A.I artists are already creating stock images that literally look like photographs you see on shutter stock, but far more specific to the prompt.

Concept art - this will be heavily affected. It includes things like landscape and environment art, character design, a lot of game-related concept art (in-game objects, user interfaces, textures, characters, etc.). Even concept art for films will be affected. I've seen so many landscape artworks, buildings, machines, and cityscapes that could easily be used for film or game concept work.

Editorial illustration - This is my field of work. The Atlantic has already published articles with A.I generated artwork. Admittedly, those particular articles may have just used stock imagery rather than an editorial illustrator's work, but it shows that art directors are aware of A.I generators and are more than willing to use them. Cosmopolitan used an A.I artist to create their cover image (although in fairness this was probably just a one-off thing). And just generally I see a lot of A.I art that could easily be used in replacement of actual editorial illustrators. I know what art directors are looking for, I know this space, I know what images work in editorial contexts and what don’t. The reality is that it will become increasingly easy for art directors to input the core idea of an article into an art generator and find a suitable image within minutes and most importantly for FREE. They don’t have to pay illustrators like me to create images. I’ve already seen people inputting music lyrics into the generators to create images. This is similar in a lot of ways to editorial work, except that editorial is just text about some political or social issue for example. It’s all the same. 

I definitely see graphic designers eventually being affected. There’s no reason why an A.I. artist can’t generate incredible website designs, logos, poster designs. It can’t do it now, but in the next 5 years I predict we will see this. You’ll be able to input a bunch of specifications, all the text and imagery you want included, and then bam! the A.I generator gives you 10 different options in less than 30 seconds. A few tweaks here and there and you’re done.

Obviously there are going to be some forms of art that won’t be affected by A.I. For example, there will always be a need and market for fine artists e.g. real paintings. Any kind of art that involves creating 'physical' stuff rather than digital work is not going to be significantly affected. 

One of the best free A.I. art generators now is called Midjourney. You can use it for free here: https://midjourney.gitbook.io/docs/

Click on the discord link, sign up, and enter /imagine [your prompt] in one of the bot groups in the left-hand side menu. 

Just seeing what Midjourney can do already is very disheartening and depressing for myself and for many other artists. 

I feel like I’m back to square one again after so many years of hard work and effort.

And like I said, I’m not saying my jobs are just going to be taken away next week. This will obviously be a slow gradual process. But the point here is that in 2, 3, or 5 years will I be needed as an artist? Or at least how many jobs will there be left? I guess the silver lining here is that no matter what I’ll always love creating art and I always can create art. No matter what I'll always have that available. 

   I can relate to the feeling and dread. I'm hard at training myself the skillsets to become a really good illustrator, then all of a sudden a A.I program can not only draw, but draw decently enough that it reminds me of how A.I learned chess or Chinese Go, they spent some time doing basic mistakes but over time learn from them to eventually becoming very good. I hate that feeling 

   At the moment, maybe using A.I generated pictures for references and further inspiration?

[LSD-Rumi]

@Leo Gura you are talking here about the top 5-10% of artists, what about people  whose job is to do simple art like logos and stuff? Their jobs will be taken for sure. No need for a human vision for a simpe logo. 

[Leo Gura]

8 minutes ago, LSD-Rumi said:

@Leo Gura you are talking here about the top 5-10% of artists, what about people  whose job is to do simple art like logos and stuff? Their jobs will be taken for sure. No need for a human vision for a simpe logo. 

1) Creating a proper and usable logo requires a lot more than some random AI image.

Random nice images != usable art for commercial applications. Commercial applications require very specific art.

2) And what about cashiers being replaced by self-checkout machines? The bottom line is that if your job can be replaced by a machine, you're not being nearly creative enough. Rather than wishing for the machines to not exist, learn to use the machines to make yourself more creative and valuable.

Edited September 1, 2022 by Leo Gura

[AtheisticNonduality]

@zurew All of those images are nonsense, non-stimulating, and even inaccurate. That's not what all those objects or subjects should look or feel like. People aren't understanding that it's completely different and cannot be otherwise. Obviously drum machines have not replaced drummers, in contexts where real drummers are needed, because they don't sound the same and don't have the same effect. That's just a minor comparison.

Take a stereotypically real piece of art, like the ceiling of the Sistine Chapel by Michelangelo, and notice that isn't something replicable by AI.

Now, maybe it could replicate the Actualized logo (though still there would have to be a human guiding it properly, which could be a job in itself), but, sure, there would be an increasing gap between artists and mere illustrators.

[Nilsi]

11 minutes ago, AtheisticNonduality said:

@zurew All of those images are nonsense, non-stimulating, and even inaccurate. That's not what all those objects or subjects should look or feel like. People aren't understanding that it's completely different and cannot be otherwise. Obviously drum machines have not replaced drummers, in contexts where real drummers are needed, because they don't sound the same and don't have the same effect. That's just a minor comparison.

Take a stereotypically real piece of art, like the ceiling of the Sistine Chapel by Michelangelo, and notice that isn't something replicable by AI.

Now, maybe it could replicate the Actualized logo (though still there would have to be a human guiding it properly, which could be a job in itself), but, sure, there would be an increasing gap between artists and mere illustrators.

I think you are underestimating AI grossly. Just because it isnt possible now (which, we dont even know what the cutting edge really is), doesnt mean it wont be possible in a year or a decade. I fully expect AI to be able to create art that is indistinguishable from the real thing, sooner rather than later. There will still need to be someone telling the AI what to dream up, but I kinda doubt that this will be something we need actual artists for. As soon as everyone has private access to some galaxy brain AI, that can literally create anything they want, they wont turn to an artist anymore. People are way too obsessed with themselves and will just create their own, personalized entertainment bubble.

[LSD-Rumi]

@AtheisticNonduality Dude what are you talking about? these AI images are absolutely amazing. These images are generated in less than 30s. Imagine what kind of technology, there will be in a decade or two. But still as Leo said, AI still requires a guider which is the human component ( the artist).

[LSD-Rumi]

@AtheisticNonduality what about those? they are not 100% perfect but still they are amazing.

[AtheisticNonduality]

18 minutes ago, LSD-Rumi said:

@AtheisticNonduality what about those? they are not 100% perfect but still they are amazing.

Neat, not exactly a place I'd want to visit though, because the details are off.

 

You are all very easily impressed. You're also underestimating the complexity of the brains of "Michelangelos" and overestimating the complexity AI researchers will reach proportionally to that.

[LSD-Rumi]

1 minute ago, AtheisticNonduality said:

You are all very easily impressed. You're also underestimating the complexity of the brains of "Michelangelos" and overestimating the complexity AI researchers will reach proportionally to that.

I think you are underestimating how fast AI develops.

[Nilsi]

51 minutes ago, AtheisticNonduality said:

You are all very easily impressed. You're also underestimating the complexity of the brains of "Michelangelos" and overestimating the complexity AI researchers will reach proportionally to that.

As far as I understand it, the AI learns that stuff itself at a certain point, and leaves the researchers in the dust, so thats no argument. I guess what Im talking about is kinda the "worst case" scenario. I hope we will have some kind of shared reality and create art for each other and hopefully AI will assist that, but I can totally see how this will lead us all into a solipsistic personalized entertainment bubble.

[Leo Gura]

Being able to generate a ton of art quickly is a very positive thing for mankind. This will just open the doors to entire worlds of entertainment.

Right now thing like games are so limited because art is so hard and expensive to produce.

Ideally an artist would never need to pick up a pencil, but just tell a computer what he wants to see.

Edited September 1, 2022 by Leo Gura

[Space]

5 hours ago, Leo Gura said:

Look, dude, artists are never gonna be replaced.

Just use the AI to help you make even better art.

This is not bad news, this great news! Humans and AI will cooperate to create amazing stuff.

You're looking at this whole thing with such a narrow perspective. There is a nearly endless need that humans have for art. Your job is not so much to put pencil to paper but to come up with art projects that are really worthwhile. Serious art projects require vision, taste, and leadership. AI will not replace that.

You have a rare opportunity here to be on the cutting edge of art, if you have some vision and play your cards right. You could makes millions and even billions here.

Imagine being able to make a graphic novel using AI by giving it the right prompts.

I appreciate your feedback and encouragement, as always. 

Just to clarify my position, I'm not saying that all human artists are going to go extinct. Obviously there will always be a need for human creativity, vision, taste, as you said.

My main point with my original post is that my particular line of work, editorial illustration, is at risk just because of the nature of the work. It will become very easy for an art director (the person who works at The Atlantic or Washington Post for example, who then works with the editorial illustrator a.k.a me) to use an art generator to find a suitable image for an article instead of contacting illustrators. 

I've already seen work that could easily be used in an editorial context. And thats with the current technology. It's not that i'm concerned about AI taking my jobs now, this week or even in the next 12 months. But soon I suspect it will start slowly infiltrating into the industry and more and more publishers (like The Atlantic, New Yorker, etc.) will start using it simple because they can get images that are good enough. 

I've spent thousands of hours working hard to become an editorial illustrator. It just feels like my entire vision for the next 5-10 years of my life is just not going to work or happen. Creating an entirely new vision, a whole new life plan is fucking hard and depressing. Particularly when it took me years to do in the first place.

I don't understand your point about making a graphic novel using AI. That wouldn't be artistic. That would just be entering prompts and flicking through the results to find what you need. Even if someone then edited the images, well you're still just working off of the AI's work. You're not coming up with the core images. But I know it's just an example to help get my mind out of this rut, so I appreciate that. 

The only path forward I have in mind at the moment is to use AI generators to produce 'idea boards', to come up with initial ideas for an illustration, which will be the basis of my actual artwork. The advantages of this are a) it would speed up my process a lot and b) i'd likely be generating higher quality/more creative ideas vs just coming up with ideas myself which can be affected by my mood, energy levels etc. So in theory I could use Midjourney as a sort of creative assistant. But again, this still doesn't detract the fact that many art directors can simply generate images themselves without needing to outsource to freelancers like myself.

2 hours ago, AtheisticNonduality said:

Take a stereotypically real piece of art, like the ceiling of the Sistine Chapel by Michelangelo, and notice that isn't something replicable by AI.

As i've mentioned in previous posts, fine artists a.k.a painters, sculptures etc, obviously won't be replaced. It's predominantly digital artists that will be affected e.g. game character design, film concept art, landscape concept art, editorial illustration, textile designers, website designers, UI design, etc etc.

Edited September 1, 2022 by Space

[Leo Gura]

12 minutes ago, Space said:

I don't understand your point about making a graphic novel using AI. That wouldn't be artistic. That would just be entering prompts and flicking through the results to find what you need.

This is where you are wrong. It is the purest art.

I want to be able to create a movie or a video game simply by telling a computer exactly what I want to see and how I want it to work. And in this case, I would be the artist. The artist is the architect of the whole thing, not the monkey who draws the lines.

You need to move up one level of abstraction. Go from illustrator to art director.

Yes, stock photos for articles will be replaced. The real question is why do you have such a limited definition of art? That's like the bottom of the barrel of the art world. I mean, at that point you might as well be painting McDonalds burger wrappers and calling it art.

The fucking Atlantic? Think bigger.

Edited September 1, 2022 by Leo Gura

[Nilsi]

This is a neat example of AI being used to create beutiful art. 

[AtheisticNonduality]

2 hours ago, LSD-Rumi said:

I think you are underestimating how fast AI develops.

Lol, I am not.

Garish and hideous and obscenely meaningless computer-generated nonsense with bright colors to lead people astray like sirens---is not art. Again, not today or in this century will there be a case of AI producing something like a Sistine Chapel ceiling, simply because reaching the complexity of the human brain this way is not practical but theoretical (and not even a good theory, just a bad theory passing for practice).