Where it used to be that creating a song was a monumental task in itself, I found myself naturally drawn towards the AI tools as I could finally independent from the dreaded relying on an independent artist of a friend or NG artist who may or may not have had time to represent the work I was looking for.
- What their thoughts are on a artwork is severely different than mines
- When you need the art is a different timeframe than theirs
- Intentionality matters
Why I say this this is because I have gotten leagues better with AI tools that it's sort of beating every expectation of what is required of an artist to be good at drafting up the AI art to tailor it to artist expression. As someone who is used to creating master-grade productions, it just seemed naive to stifle creativity since it was taking 40+ hours to draft a fully flesh-out song.
Why I say this is because with modern tools, I can look at an image like Turok 2 cover screen and do the following:
Turok 2 Cover Art
I can then process it to look like the following, which is super new but an awesome application of how to expand the image:
The image is successful rendered in an expansive 3d format that allows for you to import the image into blender. With blinder, you can use additional tools to incorporate your design to the point of it being an original image. While I would never use copyrighted images for anything (being that audio artist were already bombarded with such rules), the example follows the trend of being able to create a unique experience that is worlds different than what I would have previously experience:
This one example is a powerful reminder how independent artist though that they could match up with this level of quality, but you have to admit that this really does change the game, because of ways that current AI tools work, you can create a very carefully crafted picture out of it that looks unique:
While in this format, the drafting image process would be another step in the process, it does tell that the accuracy of the image remains well composed when it comes to layering the object for a different expression. Many photos and artworks can be rendered in a format that allows for maximum utility that came from a single .jpg or .png file, which is extremely game-changing.
It's fun, but I still think that every step of the process should be unique and original. If drafting AI artwork, every step should not include copyrighted images (save for the example above, which is purely an example). It goes to show that the artwork should be monitored for 3rd-party usage and ensuring that the compositional framework stays in the original category.
There's other examples that I can point to, but for now, I think this is a great example.