Artificial Intelligence: Helping Artists or Replacing Them?

Grace Tan '24

In the past century, the application of technology has revolutionized what humans can create. From the dizzying CGI generated patterns in Hitchcock’s “Vertigo” to the digital paint brushes on Procreate, technology has proved time and time again to increase the scope of human ability. However, such innovations propose an important question: can technology itself create art? This question has come to the forefront of discussion in the art community in recent years, especially with the rising popularity of AI-generated art. 


Modern artificial intelligence is not a new innovation, in fact, its beginnings can be traced back to 1956. However, it is only in recent years that AI tools have allowed amateurs to easily create artworks on par with pieces by professional artists. In 2022, programs such as DALL-E 2, Midjourney, and Stable Diffusion were released, making the process of creating these artworks even easier and more intuitive [1]. Whether the desired style is complex, abstract, or photorealistic, these tools can bring imagination into reality simply by employing a couple words typed into a text box. The concept of machines creating art may seem ridiculous or even impossible. However, this is exactly what AI does: give computers the ability to mimic human intelligence. 


Instead of merely following a set of rules, AI algorithms “learn” by analyzing thousands of images. Then, by referencing the aesthetics collected through analysis, AI tools are able to generate new and original artworks [2]. But how exactly does this process work?


Firstly, the text is compared with a collection of corresponding images. Although the images are encoded and thus not original, the following processes add features that make the final product completely unique. Specifically, the images go through a diffusion process that adds noise and reduces them to a sort of slush, which is then engineered into a recognizable image during the third step. Finally, the program assesses whether the product matches up with the textual input [3]. Despite this extremely technical process, it is not true that people lack any sort of role in this process. They are involved in both pre-curation and post-curation, processes in which the artist selectively sifts through inputs and outputs to obtain the desired results [2]. 


In the future, it is not unfathomable that technology will be able to create art completely independently. For now, however, even AI-generated art requires a human touch. Even if further AI development isn’t possible, the fact that machines can create original artwork is already a marvel that should be recognized.