The views of this article are the perspective of the author and may not be reflective of Confessions of the Professions.
The debate over whether human or AI is better at art is one that cannot be fully answered or understood in simple terms. It involves the fact that humans paint and draw with emotion, feelings, and human intuition and the experience of how life has treated you as a human being. A computer or any amount of lines of code can never do this, thus you can see the most beautiful painting from a human and an AI. The real question is, if another human being is looking at it, are they able to tell it was drawn by a machine?
At the start of 2023, I became fascinated with ChatGPT which led me into the art section of artificial intelligence. There were two types that of “trained computer systems” known as Dall-E 2 and Stable Diffusion, both of which come with their own unique definitions of art. How AI art actually works is that a computer processing system is trained on algorithms featuring hundreds of thousands of different images along with patterns, styles, and keywords. When a prompt is sent to this processor, the computer “draws” the image from scratch by using a technique called generative modeling.
From there, the AI uses a process called a neural network. Neural networks are inspired by the human brain and are made up of many interconnected nodes. Each node is responsible for processing a small piece of information, and the nodes are connected together in a way that allows the AI to learn patterns and relationships in the data, which then produces what appears to be an image.
There are a number of different ways to create AI art. One common method is to use a generative adversarial network (GAN). GANs are a type of machine learning algorithm that can be used to create realistic images. GANs work by training two neural networks against each other. One neural network, the generator, is responsible for creating new images. The other neural network, the discriminator, is responsible for identifying real images from fake images. The generator and discriminator are trained together, and over time, the generator becomes better at creating realistic images that the discriminator cannot distinguish from real images.
AI art is becoming amazingly good that, at times, can appear to be almost human-generated, though the art itself sometimes lacks the human emotion and specific use of color that is often involved in human-drawn or human-painted art. In the AI art world, there are a variety of different engine models popping up and websites used to generate art. MidJourney, which is on its latest version 5.1, generates a very realistic version of art that is even beyond Dall-E 2 and Stable Diffusion 2.1.
As a developer, I began to explore the realms of these types of art all in a single place and with the help of ChatGPT 3, I created artsy.sh, which at first tended to generate oddly weird shaped art, but as my prompts began to sharpen, I could generate art beyond words. The majority of images you see in this article were all generated with artsy.sh. Artsy is completely free and tries to encapsulate the best of art from all different engine models into a single place. With it, I created the following images:
While AI art is becoming extremely controversial in the real world, especially in art competitions, it is still winning many awards from all over. AI art has yet to be banned, and real artists who spend hours painting are furious their works are being overlooked, but it is still awe-inspiring to observers. There is something in the art of the prompt, with many new “prompt engineering jobs” springing up with companies willing to pay 6-figure digits to engineer the most unimaginable art possible.
From vintage to traditional to modern, the art business is one worth billions of dollars from real artists. The AI-generated art is likely to also drawn in billions of dollars with people displaying their art. Whether it is regulated and artists are required to depict whether it was generated by art or not is still completely open, as no laws nor art-world entities moving to ban AI art just yet. However, more artists are opting to to display banners on whether their art was generated by AI or not.
There is no doubt that even the most precious unique prompt can depict some of the most precious art. The wondrous nature of AI art is the fact that there are many people who are not artists, cannot paint to save their lives, and yet, have found comfort in the fact that with a single detailed prompt can produce art in so many different ways.
The truth is that AI art is still in its early days of exploration and discovery. It still a controversial topic, but it is very likely AI art is only going to get more advanced and someday even include “the art of human emotion”. Mastering the skill of AI art does not have to be complicated, but becomes better as machine learning is evolving. The beauty of AI art is that anyone can do it, whether you can draw or paint or not. All you need to do is have some understanding of the English language, and you too, can generate images beyond your imagination.
Get started today with artsy.sh!