After my last deep-dive into seeing how I could utilize Midjourney to create letter-forms, I switched to try another tool and tested the newly-released Adobe Firefly. Results using this were more promising, especially when uploading a reference image.
From those experiments it became apparent, that the key to achieving good results with AI is to control the input i.e. training data. The current on-demand software solutions, however, allow for this to be done only in a very limited way – that is by uploading a reference image. As an alternative to this, it is possible to train an open-source AI by oneself. Stable Diffusion is an AI developed by the CompVis Group from the LMU Munich and Runway, and it is one of the deep learning models that can be trained by end-users. I was intrigued by this and decided to give it a try. Unfortunately, after setting up the AI and using DreamBooth to fine-tune it, I was not successful in completing the training of the model.
It appears to me, that when trying to utilize AI to create and set type, the area is still very much in its infancy. While functional ready-made solutions like Midjourney and Adobe Firefly exist for general image-creation, none of them work very well for typography yet. Training an AI oneself seems to require an excessive knowledge of the programming language Python and a rather expensive set-up of equipment.
After this realization, I decided to take a step back and look at my topic again. Should I change it or open my perspective? After some further research and helpful conversations, I came to the conclusion that I am going to change my topic slightly and open it up to research future typographic developments in general. My key question that I would like to follow is…
“What will typography look like in the future?”
I want to tackle the topic from a variety of different perspectives, making forecasts/prognoses for future developments from each perspective in order to get a holistic picture of how the typographic scenes and the practices designers engage with might change in the near future.
Possible perspectives could be:
The future of type is…
- Variable?
- Kinetic?
- Non-binary / feminist?
- Inclusive?
- Multilingual / multi-script?
- Sustainable?
- Expressive?
- Made with artificial intelligence?
I already looked into the last perspective. In the upcoming days I will begin to research the remaining ones and add to this list.
