The Future of Illustration
Agnes B. Muren & Jenny K. G. Paus
Recently, programs like Midjourney and DALL-E 2 have taken the world by storm with their use of AI to generate images
using simple textual prompts. Easily accessed images are generated in mere seconds that only get more and more accurate,
according to the human prompter, as time goes by. How will this affect visual communication designers, image makers and
graphic design? And can we, as a profession, learn to work with AI, and not against it?
In this article we asked for opinions on this subject from illustrator Gunnvor Rasmussen and comic book creator Jens K. Styve in the hope of getting answers to these questions.
Gunvor Rasmussen works as an illustrator and owns her own gallery on Bryggen in Bergen, Norway. She sells handmade illustrations made with ink and watercolour, illustrated books, t-shirts, postcards, prints, and more. She had her debut in the newspaper Bergens Tidene in 2006 when she was studying Visual Communication at KHIB.
Jens K. Styve is a comic book artist based in Tromsø. His comic, Dunce, is published daily in the newspaper Klassekampen and sixteen other Norwegian newspapers. It came first in the category for best Norwegian comic in 2019.
Styve is mostly positive about AI and believes that it is an exciting and inspiring opportunity to explore AI-based
image generators, such as Midjourney, even though he can clearly see how it could replace an illustrator in many cases.
He says, ‘I even knew a client that used Midjourney to make the frontpage of a catalogue, where he normally would have hired an illustrator, AI is clearly an immediate “threat” for an illustrator in some jobs.’
The fear of being replaced by machines has been there since the Industrial Revolution, Rasmussen says. And she continues to say, new technology can be fearful for many specialized in the field. ‘To be branded redundant and outdated can feel threatening, no one wants that. So even though changes can be very healthy, it can feel completely opposite for the artform.’
Styve has become more of a comic book creator than an illustrator, and as for comics created by AI he has thus far not seen anything useful. Comics are characterised by a series of pictures that tell a story. And as Jens himself says: ‘Comics are largely about creating stories; the day AI replaces comics it might as well replace novels and screenplays.’ Which we have found out, through our collective explorations of AI for Ymt 5, that it just may do—soon, if not already—or at least give human authors something to worry about or play with depending.
Rasmussen has done a few tests in Midjourney and DALL-E 2, where she explored whether she could replicate her own illustrative style in ink and watercolour; she concluded that was not something the AI managed to do, but Rasmussen was still able to use what she created as ideas for sketching and for her inspiration. She owns the creative license to her own illustration style, and as she puts it herself: ‘Commissions that come my way I usually get because it is my style the customer wants, and they contact me because this is a job, they don't have specialised knowledge to make themselves. It is a little bit like how you hire an accountant or plumber.’
Instead of AI being something that replaces you, can it possibly be put to use as a tool to help us create?
Styve have already experimented with how he can use AI for sketching. He has asked Midjourney to draw certain types of landscapes or buildings. He then uses the generated pictures as reference and a base for when he starts to sketch. Having a vast collection of references at your fingertips can be a great resource to have as an artist. As Rasmussen mentioned: ‘I rather think that Stockphoto will get big competition.’
She has seen some amazing things be made by people who are talented, and she explains that they spent a lot of time to illustrate just what they imagined. The images of quality that Midjourney generates are often promted by designers and artists or others in the creative field. They are the AI users who understand the terms one must know for the generated work to become something good—they have enough knowledge to base their prompts to the AI on.