The first invite that I lodged in Dall-E 2 was “the cubist painting of a tuxedo cat playing the piano”. The AI tinkered and stamped for a minute, then delivered four absolutely adorable images of cats playing the piano. The surprise, the fantasy and the color impressed me – after all, many of us have seen some of the More disturbing results images generated by AI.
I teach and wrote and bite with comics for a few years now, so it seemed natural to experience Dall-E 2 to make a text image collaboration. Would Dall-E 2 and I develop a unique and unique “voice” in our creative outing?
As a rule, I make comics in two ways: I draw and write something quickly in a newspaper where I spend months script, write and rewrite, draw and erase, digitize and edit half a dozen pages. Makeing bands is a laborious process, especially for the comics of a unique creator. In this “author” model of comics manufacturing, the person who writes comics also illustrates it. From Underground Comix (Transgressive comics that did not respond to the approval of the comic book Authority code) and alternative comics (alternative to, say, dominant Marvel and DC comics), these hand -drawn comics reflect the unique and developed style of their creator. Beautiful brush of Craig Thompson has John PorcellinoThe clean lines and the distilled images, the panels of a creator of comics become as recognizable as the voice of a writer or an artist style.
Caricaturist Lynda Barry defends the beauty of the hand drawn line and the individuality of her expression. Even if you are “bad in drawing”, you make a unique image to you; Barry would say that your drawing is more “alive” than the drawing of a professional designer. There is, she argues, something personal and real about an inaccurate representation drawing. Even the physical appearance of the text becomes another method of communication, a kind of image, because handwriting transports an innate expression of personality and individuality, offering another layer of tone and meaning.
Some of the Dall-E 2 drawings have part of this “living” quality. After downloading one of my own sketches of a playful cat in a witch hat, Dall-E 2 made four variations which, for me, seemed to be “inhabited” characters. But it is difficult to regularly reproduce these qualities with a new prompt (to be fair, it is also a challenge for me).
A complete collaboration between two AI to “generate” a comic strip is now possible that Optai launched Cat And InstructgptWhich generates writing when you invite it. But the comics are based on a poetic precision of the line – there is not much space for the explanatory text, and I like to prevent the words from duplicating the work that images do.
Describing the image I wanted the book to have taken another type of finesse and revision to which I was used to. The fast box reminded me that longer and more specific descriptions will give better results. Therefore, the writing work is doubled, scripting comics and describing the desired possible illustrations. “The pastel drawing in the chalk of a night sky with many stars” turned out to be a disappointment but generated a new idea: “pastel drawing in the chalk of an equestrian farm in Kentucky at night with a lot of stars in the sky.” Have I trained the AI? Did the AI trained me? Or have we engaged in an iterative process as employees? I finished the experience, asking myself what it might need to make a complete collaboration in which the images completed writing and felt expressive, “living” rather than soulless.
