BY EDWARD STEURY
STAFF WRITER
AI-generated art is a strange and wonderful field, and although some are concerned about its impacts on art and the expression of creative freedom, I am here to somewhat naively sway that belief. To begin, I should perhaps explain the brief history of AI-generated art. The first real boom in public interest came from DALLE Mini, a small program that would generate nine blurry, incomprehensible piles of visual noise that sometimes vaguely resembled what you asked for. To me, interest was only gained because people looked upon DALLE Mini as a joke or because they were worried. After all, they supposed, “That blob of red kind of looks like a solo cup! That’s what I asked for! THIS IS THE TERMINATOR TIMELINE!” But then, when interest died down, as it so often does on the Internet, the topic of today’s article was born.
Stable diffusion was a new model of AI art generation that took the community by storm. It was fast-delivered, defined, and high-definition images with a small amount of cohesion. Below are two attempted recreations of our school’s mascots: the MSA Dragon and the Fighting Calculator.
Impressive, especially considering that a few months ago, people were gawking at DALLE Mini–clumps of data congealing into Google’s top image results. However, the drawbacks of stable diffusion are pretty evident. To get anywhere near the masterpieces that dot the top AI-generated pieces, the programs require an enormous amount of memory. And if you’re going to make a server that processes all these requests, that is going to cost a lot of money. So that’s why all the top stable diffusion websites–DALLE 2, Midjourney, and Dreamstudio–all have some form of expensive payment or credit plan. Oftentimes, a hundred generations will run you around ten dollars; sometimes, it’s even pricier. On the other end of the spectrum, a program like stable diffusion on a home computer will cost you an arm and a leg when it comes to RAM.
However, I went into this promising to tackle the difficult moral question of is AI art generation ethical? I would argue that it is because AI can help improve so many aspects of the creation process, such as inspiration, difficult and tedious processes, and more. Plus, let’s be honest, my examples above aren’t exactly masterpieces, and AI art generation still needs to base its model on real art or real photos. So, in conclusion, don’t worry about robots taking over art. Instead, think of it more as a supplement to the field, because isn’t art about seeing things from different angles? And what’s a more artistically interesting angle than that of a robot?