The project I would like to share is DALL·E, which is a machine-learning program that generates images based on natural language description. DALL·E has been insanely popularized these recent months by artists who use it to create “AI art”.
There is a large room for improvement in the future of DALL·E. Currently, because of how the algorithm “mush together” found online images, generated images usually have less defined edges and distorted features. They appear more abstract and like a “fever dream”, where objects all resemble something but the details are all wrong. This is part of the reason why artists now mostly use it to generate abstract concept art or fix up the images later by drawing. However, with better algorithms implemented it could generate much more precise images in the future.
DALL·E has sparked heated discussions among traditional as well as digital artists about whether or not “machine will overtake human artists”. This artificial intelligence is revolutionizing the art industry drastically and forcing artists to rethink the definition, meaning, and scope of art. It reminds me of how the invention of photography caused revolution and panic in the traditional art world. DALL·E, and many other AI based image generator, should not be feared in the art world. They mark a turning point in the art-related technology. Now we can turn the images in our heads to image by a click of a button, what’s next?
