I spent a lot of my leisure time over the summer vacation producing paintings. I don’t mean with a paintbrush or some colored pencils, but with Dream, an app on my phone. Dream, which was made by the Canadian company WOMBO, which is known for its deepfakes with lip-syncing, lets users “create stunning art with the power of AI.” Enter a concept in 100 characters or fewer, choose an art style, and the app will generate a picture in seconds based on the prompt.
Dream originally emerged on the App Store in late 2021, and it has since become a social media sensation. There are TikToks where people ask for photos based on the titles of books, and other people have made a “game” where they put the names of celebrities into the app and have viewers guess who it is based on how similar it looks to them. Some Twitter users have even developed accounts devoted to exhibiting “dream art,” such as this one for Taylor Swift lyrics (a personal favorite). The images circulating on the internet are charmingly odd, but they capture their subjects.
Dream is not the first of its sort; AI art has been around for a long time. In 2018, a portrait created using GAN technology sold for $432,000 USD. Many experts and critics believe that AI will not replace traditional art anytime soon since most computer-generated creations are just copies of existing patterns and structures. Instead, the power of AI art comes in human cooperation. Take, for example, South Korean artist Domin, who collaborated with an AI artist for a show at the AI.A Art Gallery. Domin created part of a Dokdo Island scenario — one with a pen, the other with a paintbrush — and the AI finished the image. Both designs were then linked together, fusing not just Eastern and Western art forms, but also human art and computer graphics.
Dream is a means to make art more accessible, both technologically and artistically. It’s astounding to think that all of this work is done using algorithms (WOMBO has been tight-lipped on how the app works, although it’s thought to employ VQGAN+CLIP technology.) I don’t claim to understand how AI works beyond the basics, but it doesn’t diminish my awe for such a feat. From start to finish, building and deploying AI might cost up to $1,000,000, yet Dream is free to use (although purchasing larger prints of your artwork is an option through the app.)
To put it simply, not everyone is born with a natural flair for classical painting. While entering a few lines and letting software handle the rest is not comparable to the hours of time and effort required to create a complex painting, this medium broadens horizons and enables everyone interested in digital art to try.
One of the reasons I like the program is the degree of intricacy it provides. Creators of movies or music compositions have a flair for presenting a narrative with the smallest details: a little shift in facial expression or a developing crescendo. At first sight, I find it more difficult to see such details in actual works of art. Dream artwork effectively captures complicated thoughts and feelings in a more apparent form – for example, I put the title of this post through the program to produce the cover picture. The algorithms have managed to precisely represent this idea: the beauty that may come from turning digital. It’s amazing to see something so complex come to life so soon.
AI art may not ever equal human brilliance, but it is definitely here to stay. Inventions like “Dream” make technology accessible to the public and foster a deeper appreciation for art in general.