Many artists have died, but DALL-E is capable of amplifying their works of art. This is how they would look according to the AI
DALL-E continues to surprise us every day. We don’t know how far this AI will go, but users are managing to create authentic works of art in a matter of seconds. The latest trend among the community is zoom out on the great works of art in history. Although the original authors never drew or commented on what lay beyond their most famous works, the AI manages to enlarge these paintings with impressive quality.
From the Reddit forum we found some of the images that we show you below. They are just a small sample of what DALL-E 2 is capable of representing. Carlos Santana, an AI promoter on YouTube, also offers several of the examples that we have shared here through its Twitter profile. The question lies in choose the artwork and ask DALL-E to fill in a larger spacefollowing the idea of the main work.
From “La Luna Mona” to the extended Guernica
In the cover image we find “La Luna Mona”, as Santana has named. It is the famous painting of the Mona Lisa but with a greater horizontal space, where the AI has continued the mountains and has decided to place a moon in the upper right corner.
In the image above we find Van Gogh’s famous painting of The Starry Night, but where the scene is enlarged to represent a city canal and the darkest space in the stars.
The seagull and the man. pic.twitter.com/1M3dI3410G
— Carlos Santana (@DotCSV) July 7, 2022
The selected images are some of the options generated by DALL-E, since we remember that the AI offers various results to the user.
Only by selecting a smaller box or area is it possible to enlarge the image, with no size limit. Yes ok, as we ask the AI to fill in more space, the synchronization with the original decreases. After all, these great works of art contain a very precise composition that is difficult to modify.
It is also striking to see how the AI decides implement objects that were not in the original box, but who understands that they fit with the idea of the painting. This is the case of the moon with the work of Da Vinci, or with a cherry tree in bloom next to the famous Japanese wave.
The possibilities are enormous. All you need is a good ‘prompt’ (the phrase that we indicate to the AI to generate the image). Enlarging artwork is not new and many artists have done the same in the past. With DALL-E 2 all this is done much faster. Instead of having to imagine it ourselves and recreate it, for example with Photoshop, now it is the AI that allows you to do it faster. And with ideas that do not have to be ours.
What Alba CeColl points outart historian and creator of digital content: “the narrative force of the painting within the painting is lost, by extending the work unnecessarily. Most of the compositions that DALLE2 had made they took away all compositional force from the original work“. You are right that AI still has a long way to go, but it does not cease to amaze (once again) the possibilities in art that have been opened up thanks to the arrival of DALL-E and the like.
In Xataka | DALL-E works with the images of creators who do not receive anything in return: what copyright says about AI
