The Lines Between AI and Art are Blurrier Than You Think

You can hardly open your facebook or instagram today that you are not blasted with the AI generated images.  Some are pretty good; some are just ridiculous and hardly worth a glance as you scroll past.  I personally was adamantly against anything that even hinted of AI generation when the movement started.  BUT.. my opinion had changed! 

I’m going to explain why, show you a piece of art that I produced using some AI generated elements and briefly discuss why I feel the lines between art compositing resources like stock and AI are now so blurred, it is hard to differentiate what you are working with.

First all, why did my opinion change?  Well, I got educated on the process.  I took a course on how the AI engines actually work.  I now understand that AI engines don’t actually “steal” or “store” other artists work somewhere in a large database and use their work without their permission, but rather they use computer algorithms to “learning” based on being present hundreds of thousands, even millions of samples of images that they can use for recall and recompile an image based on the word prompts you feed it.  This process isn’t radically different than how a human forms concepts from those visual experiences that we have stored in your little digital computer called our brain.   The AI engines are only as good and the programmers who build and enhance them. 

If you would like to learn more about the Midjourney AI process and how to get started with Discord  you can get a great class here from Phil Steele.  I think he is currently running a 40% off special so don’t wait to sign up.  

But that doesn’t mean that there are no ethical considerations for what you prompt and get back from these tools.  But I will leave that discussion for another time. 

Today, I want to talk about the elements you get for your art composites and how these elements are not much different from stock sites when you prompt the AI engines properly.  As an artist and compositor, I’m interested in the AI engine producing more photorealistic and detailed images, as well as elements isolated on white backgrounds or scenic backgrounds and settings that I can use in my art.  

So here is the first piece of art that I created using elements I generated or downloaded from Midjourney.  It’s called the Queen of Venus.

The image I originally got back from the AI engine was this.

I wasn’t happy with the model’s face, even though I had prompted for photo realistic, ultra detailed and several other photo details.  So I went to my library of model images and selected a model that I could take just the facial triangle  (eyes, nose and mouth) and insert it into my composite.  Here is the model that I used.   

I then selected the AI Venus off the blue background, adjust the colors and built my own nebula background using several of my own stock images of the cosmos I already had in my library.  

 

Here are the Cosmos images I used

I wanted to add elements such as a planetary landscape and some planets so I returned to Midjourney.  But instead of me generating these elements by feeding the engine prompts, I used the “explore” feature and just asked to see what others had already built that I could use.  I asked for “planets on white background” and with a second request I asked for a “landscape of Mars”.  I got several new elements to add to my stock library but these three were used in this composite.

The remainder of my composite is made up of splash and cloud brushes, particle brushes and various textures and colored LUTS.

My point of showing you the elements in this compositions is this.  THESE PLANETS and MARS LANDSCAPE images, while generated in Midjourney could just as well have been downloaded from Shutterstock, Adobe Stock or any other stock site. Some 3d artist could have built them with blender or Das 3d or Maya and sold them to the stock site.  Some artist could have generated them in one of the many AI engines and added them to their stock libraries because most of the large stock vendors now accept AI generated images!

If you are creating art and you want to be a purist and not use ANYTHING AI, then you better be a great photographer and be committed to capturing all your own content.   Because today, if you are visiting stock sites, you really have NO GUARENTEE of the source of materials that you download.  Some stock sites are stating that the image is AI generated – BUT WHO REALLY CARES.  DOES it matter if it was built with Das 3d or Midjourney?

The real debate is not AI or not, but YOUR OWN INTEGRITY in what YOU PRODUCE AND CALL YOUR ART!! 

If you are generating something with simple word prompts and putting it back out into the public domain and calling it YOUR ART… than shame on you!  That is NOT ART!   That is a testament to the technical programmers that built the AI engine!!

If you use the AI engines as a tool to produce concepts or images for inspiration, or generate results to help you visualize a scene for cinematic lighting or shadows to improve your composites, or you build elements for selections to add or create your composites than you are using AI as an artistic tool.  To call it your art… you still have to BE THE ARTIST!

 

 


Galleries:  The Pixel Mixer.  

My Stock Profiles:  Shutterstock.   Adobe Stock.   Dreamstime

#digitalart #art #artist #artwork #fanart #digitaldrawing #digitalillustration #digitalpainting #digitalartist #painting #photoshop #digitalartwork #compositing #photomanipulation #creative #photoediting #photoshopart #digitalmanipulation #prints

 

Comments are closed.