The first time I really thought about the role artificial intelligence could play in design was when mentoring Jean Böhms Master Thesis Pathfinder in 2020 at the University of Mainz. While being his tutor and auditor focusing on the formalities of scientific investigation and writing, I have been secretly his student, learning so much from him about the history and possible future of computational systems in type design. Exciting times seemed to lie ahead of us. So much of the tedious work that is done at the moment by experts, seemed to be taken over by artificial intelligence any time soon. After all, type design is one of the oldest crafts of design and there are millions of examples of legible fonts of any type that can be used to build datasets. Even if we would not want to automate the design of the entire font, AI could help us imagine the rest of the font, when only showing it a couple of letters we have designed. Prediction, based on datasets, is one of AI’s strong suits. Very exciting!

Two years later, in 2022, there was no escape from mediocre design, illustrations, and texts “created” by A.I. in social media anymore. The excitement for the NEW had evaporated. All that was left was millions of people doing the same boring stuff, probably because only a few built their own datasets and trained the algorithm themselves. This means that they were all feeding off the same stolen content and using the same algorithms mixing up something that has already been created by humans.

I was not just bored of images created by A.I. I was concerned as an educator. I was taught by a swiss-canadian typography teacher, Prof. Dr. Sandra Hoffmann Robbiani, who has studied and taught in Basel, one of the best schools for typography that has ever existed, and how things are going, it will probably stay that way. One of the main teaching methods has been familiarizing yourself with the subtleness of design through physical approximation. I will give you an example. Instead of telling us how to choose the right font size, line spacing, and width of the column, and set a proper justified or left-aligned text, we were just told to do so. I studied design in the late 1990s, which means we had computers around, but they did not serve educational purpose. We were told to go the much harder way and do this by hand. It was a mess. We copied or printed the text, cut out each word with a scalpel, and glue them back on a piece of paper. It took us weeks just to get the lines straight and learn how to glue without making everything dirty. We were all frustrated and asked ourselves how learning to glue something is going to help us in a job where nothing is glued and all is done on the computer. Much later I understood that slowing us down, helping us to pay attention to detail, and forcing us to do it by hand, enabled us to see and feel things we haven’t before, which is a huge difference from just knowing.

In other words, if we outsource the creative process, we also outsource thinking. Making is deeply connected with thinking. If you want to learn or develop something new, you have to go through a painful process if you want it to be worth something. Take Writing. Writing is an extremely painful process if you want to come up with something new, because you need every cell of your brain to manifest your thoughts on a paper and make them readable for someone else, and by doing so, you are making them readable to yourself. Your text is becoming the sparring partner of your thought process. Automating the writing process or any kind of creative process (through A.I.) robs you of an important part of thinking, the externalization of an idea, and the consequential opportunity to evaluate it from different perspectives.

You might think that I am against A.I. Not really. There are tons of applications where it can help us to reach goals that were out of reach before and do boring tasks nobody liked to do anyway. I am neither for A.I. nor against it. We have invented it, and now we have to find smart ways to deal with it. It is naive to think we can avoid evolution. (See Red Queen Hypothesis) Dealing with it smartly means first properly understanding it and then asking what we want to use it for and if it is worth using tons of non-renewable-in-our-lifetime energy and materials for it. Even if we would find ways to make A.I. more energy efficient and energy renewable, we would probably consume more anyway, because we can. (See Jevons Paradox and Moore’s Law) But again, that seems to be part of our evolution too.

In economics, the Jevons paradox occurs when technological progress or government policy increases the efficiency with which a resource is used (reducing the amount necessary for any one use), but the falling cost of use increases its demand, increasing, rather than reducing, resource use.
https://en.wikipedia.org/wiki/Jevons_paradox

First I gave a lecture about what A.I. is, how it is currently used, and what it could be used for, focusing on design. Then I took them on a journey to the history of modular type design and establishing the required ingredients. I showed them what kind of grid produces what kind of design and what modules we need to at least draw uppercase letters.

Then we learned which prompts produced the most interesting results and how we, as a group, use A.I. to produce our modules. Learning to prompt properly is useful when moving to conversational UI. It was important to me that we shared the modules A.I. produced, to provoke a reflection on the definition of ownership in the context of working with datasets that we do not have ownership over. We produced 745 modules in only one day.

The following steps required human intelligence. How would the students turn the modules into typeable fonts? Which is a very different mindset, a more systemic mindset, than designing letters for lettering.

At the end of the workshop, every student had to design and print a poster with Riso, emphasizing the importance of the nuances of human communication. Not just by enabling human decision-making in the design process, but as well as the physical experience of a special printing technique, the Riso printer.