How Neural Networks such as “Stable Diffusion” will influence the World of Art

How neural networks such as “Stable Diffusion” will influence the world of art?

I think the future of freelancers and just artists who paint “for peace of mind” is rather sad as long as capitalism exists, corporations concentrate all power and all production under their wing, fire employees, and leave only directors, managers, and accounting. Or maybe there will be new vacancies for “image quality control”, which will bring the world’s bureaucracy to the level of absolutism.

What does it feel like some kind of neural network can create 5,000 images in a second that didn’t exist before? No need for artists, no education, the whole essence of art can come to naught.

Despite this, there is something to be glad about, progress does not stand still, and let’s hope that neural networks will benefit humanity and ultimately lead to further development, but not deplorable events.

The neural network can generate pictures in any style, everything that exists and is present on the Internet. Dall*e and Stable diffusion works in the “find and copy” format, you set the parameters, for example, “beautiful young fantasy princess, diffuse lighting, fantasy, intricate, elegant, highly detailed, lifelike, photorealistic, digital painting, artstation, illustration, concept art, smooth, sharp focus, art by john collier and albert aublet and krenz cushart and artem demura and alphonse mucha” and the neural network gives you the results.

Specifying what you want to draw, in what style, and which artists you want to use as an example for your AI. Despite this, the pictures come out unique (Even Google Lense does not see any catch).

However, not all pictures come out perfect, neural networks still have a lot to learn, especially to draw pupils or hands, as well as letters. Due to the presence of a huge variety of fonts, languages, writing styles, the neural network simply cannot figure out what to choose, and as a result, “some kind of dregs” may result.

The neural network can generate pictures in any style, everything that exists and is present on the Internet. Dalle and Stable Diffusion works in the “find and copy” format, you set the parameters, for example, “beautiful young fantasy princess, diffuse lighting, fantasy, intricate, elegant, highly detailed, lifelike, photorealistic, digital painting, artstation, illustration, concept art, smooth, sharp focus, art by John Collier and Albert Aublet and Krenz Cushart and Artem Demura and Alphonse Mucha” (AI still draws hands the wrong way):

You can specify what you want to draw, in what style, and which artists you want to use as an example for your AI. Despite this, the pictures come out unique (Even Google Lense does not see any catch).

However, not all pictures come out perfect, neural networks still have a lot to learn, especially to draw pupils or hands, as well as letters. Due to the presence of a huge variety of fonts, languages, writing styles, the neural network simply cannot figure out what to choose, and as a result, “some kind of dregs” may result.

Yeah, it is sad people may be replaced even in art, and yeah, AI draws additional fingers, weird faces and sometimes pictures look like consequences of a nuclear war but it still DRAWS! 😀

We have prepared a selection of pictures for you, guess what was drawn by a neural network and what was drawn by a real artist. Write in the comments what in your opinion is the work of the artist, and what was created using the “Stable Diffusion” neural network.

Stable Diffusion AI or Real Artist?

PiSi: All pictures here are generated by Stable Diffusion AI

If you have recognized ALL ARTS here as AI art, then you may be an advanced artist or higher, or just have good sense or taste.