A.I. Generated Art: Will it Replace Artists in the Future?

By Ricardo He, 12th Grade

On August 30^th of this year, a man named Jason Allen won first place in Colorado’s State Fair fine arts competition utilizing a combination of A.I. software Midjourney and Gigapixel. The triumph of the A.I. generated artwork over human-made artwork caused an uproar in the art community. Many of them attacking and criticizing Allen, declaring that he had cheated and had put no effort into making art. Many have also shared the opinion of a Twitter user stating that “typing keywords in a good sequence isn’t art”.

Although A.I. generated art has been in development since the 1960s, Harold Cohen’s AARON being one of the first, this AI generator could create infinite images, however, it had limitations. AARON could only create images using the system’s own style. Nowadays, these limitations are starting to crumble. With the rise of the two of the most popular software at the time of writing, DALL-E 2 and Midjourney use text-based prompts to generate images and are capable of learning different styles. For example, take a look at the following portrait painting by Picasso.

Now let’s compare it to an AI-generated image imitating Picasso’s style.

Now, what if I told you that the roles were inverted and that the first image was completely AI generated while the second one was Picasso’s Portrait of Olga? Would you believe it? Even if you managed to spot the fake, you could definitely see the similarities in style between the real artist and the AI-generated image. Picasso was famous for creating cubism, which rejected the standard concept that art should copy nature, and utilized basic geometric forms to portray the real world. The image created by the AI quite noticeably reproduced this type of painting in mere seconds.

Let’s try another one, this time imitating Van Gogh’s art style.

The prompt utilized for this image was: “A painting of a starry night sky with the London tower and a small astronaut in the sky utilizing Van Gogh’s art style”. Although the similarities are “less” obvious compared to that of imitating Picasso’s art style, it is still evident that the software is coming close to replicating Van Gogh’s Post-Impressionism art style; that is, an art style that adopted the use of vivid but arbitrary colors with distinctive brush strokes that emphasized geometric forms. Moreover, let’s not forget that the software DALL-E 2 is still in its Beta stage, meaning that given some time, it will surely become better and better.

DALL-E 2 is a free software available for anyone right now. You sign up and get 50 free prompts, which later you will have to pay to keep using the software. Midjourney is almost identical, the difference being that Midjourney gives you 25 free prompts before having to buy more. Taking this into account, how will the availability of this software and others affect the future of traditional artists?

Well, with software being able to freely generate almost precise and high-quality images, it is undeniable that artists will definitely experience adversity, having a harder time getting customers; especially the smaller, less experienced artists, as their job can be easily done by these machines. Meaning that smaller companies are on a tight budget and the average person looking for a quick painting or digital image would go for the cheap (or even free) and quick option. However, to say that an artist’s job will completely disappear because of AI is an absurd statement, as every so often, words cannot and will not be able to describe precisely what a person is thinking of, or wants to see on the canvas, and that is where human-made art will be needed.

References: