Random Spiel, It's Not the End for Proving An Art Process
I don't think there's going back now that there are AI tools for generating images for art. So rather than spend time denying how this affects my interests, I thought it best to understand the subject better than dismiss new technological possibilities.
Just like a time when portrait painters saw the invention of the camera as a threat to their livelihoods, I think this one of those periods when art reshapes itself to suit the times. Art will eventually adapt and it may not be at a direction the popular opinion agrees with for now. I'll just say stuff based on my own plebian opinion and let the art experts work it out.
Below is a video that inspired me to write this post from the AISearch YT Channel. It's a cool channel that features new AI tools and the latest news about AI. I watch these types of content. Below is a video with instructions on how to setup an AI that produces generated images and videos like time lapsed or process shots.
It can mimic those art contents on youtube that features artist time lapsed videos showing their art process and also give the user the final output. It's like a fake artist starter pack except now that there's AI bros that treat these generated images as art.
In a previous post about Runway Gen3 can fake art process videos implies this is just the beginner of a series of tools that will try to get better at selling fake artist profiles online. The witch hunts happening among artists are also an existing problem where artists accuse other artists of having AI generated images but it's actually a difference in skill issue and experience. It's easy to accuse someone of being a fake without proof and somehow the burden of proof lies in the artist having their name dragged to the mud somehow.
I can see the flaws in the process videos that can escape beginner artists. I'm still considering myself as a beginner but I'm several hours deep when it comes to consuming time lapsed videos and process shots so when I saw how the tool generates the time lapse videos, there are several artifacts that in the images that can easily be missed for beginner artists. The sad part about this is that it takes a few more looks before the pattern becomes noticeable and bystanders who just see pictures they love press like would probably miss the detail.
The part that grinds my gears when it comes to using AI art to do self marketing as an artist is the failure of disclosing you're using AI generated images or used a bit of AI on your work while framing it like you did everything old school. It's purposely trying to omit a crucial detail where some people want nothing to do with AI generate images. Whether you use AI or not, you do you but be genuine about your process and not fool people.
Anyway, the tool above can also be used to break down parts for study. Let's say you saw an a piece and want to breakdown the details to "study" the parts and how to render, the AI tool in the video above shows how it breaks down the image from line art to colors so that you can follow. This is one of the features of this tool I can at least get behind because it's a function that helps an artist "study" to better themselves. But I know there are people out there that don't limit their imagination to this application and can also deconstruct and reconstruct the piece to suit their own generated image version to sell on some Patreon account.
Here's a major flaw this tool has yet to overcome is showing the UI for the art program used during the process. It only shows the workspace but not the surrounding icons of known art programs like Clip Studio, Procreate, Photoshop and etc which a lot of artists also customize their field. The images also lack the brush works and intended effects making the artifacts more visible. That part requires a little bit of experience when it comes to art curation to weed out the fakes.
It's not the end for proving your art process if you're into creating time lapsed videos or just settling for screen shots. A human artist can explain why they rendered something in the piece that way while the generated image will show details that don't make sense. It all boils down to intentionally rendering a detail that way versus letting the text prompt and parameters take the wheel.
Thanks for your time.
Posted Using InLeo Alpha
There’s definitely a lot of things we will need to learn how to deal with and manage on the AI front.