We’ve seen the news about DALL-E and MidJourney. Type in a few keywords about anything and everything, and you’ll get a generated image that does not disappoint, and even surprises.
Haven’t heard of them yet? These “AI” image generators are essentially machine learning models and artificial intelligence programs that generate digital images from textual language prompts (or natural language descriptions).
The Economist’s June 2022 Issue Cover
To start, you type a command ‘/imagine’ and a few keywords such as ‘a chicken painting in Chinese calligraphy style’, and the MidJourney bot on Discord will generate a set of images.
/imagine ‘a chicken painting in Chinese calligraphy style’
Next, you can choose ‘Upscale’ or ‘Variation’ with each image.
Generated image, upscaled
Fancy something else? Try something different such as ‘multiple faces in one head, pop art Andy Warhol and cybernetic’.
/imagine ‘multiple faces in one head, pop art andy warhol and cybernetic’
Generated image with selected variations
Its ‘imaginative’ abilities are expansive. From a world of traditional art techniques to the style of Andy Warhol’s Pop Art, Andreas Gursky’s photographs, and anything you can imagine, this AI bot does it well, efficiently and with surprising artistic flair.
Artists were once thought to be completely immune to job automation. And indeed, such AI programs will change the way we approach and even necessitate the need for artists, photographers, designers and probably many more. Just imagine the disruption when MidJourney or Dall-E adds the option to generate a video or animation in addition to static pictures…
But before the creative world completely shuns such developments due to fear-mongering, pioneer Computational Artist Frieder Nake has something to say about the intelligence of AI.
He concludes the role of the artist succinctly:
This is the end of Part 1. In Part 2: Artist and AI, we explore its implications on copyright and more.