Latent space is a space created within machine learning processes. Objects in this space will be positioned depending on different resemblances defined by variables which results in the emergence of a space with many dimensions, connections and networks.
We decided to use the time in our residency to explore the topic of artificial intelligence (AI) aka machine learning – hence to make our first steps into latent space. But let us start from the beginning.
Last year we heard about AI’s being able to create images based on natural language and it was an intriguing vision. Being able to just type a sentence that comes to mind and shortly after being presented an image (or even a variety of images to pick from). We heard about it from the distance sometimes saw some AI generated images—but never tried it by ourselves. So coming to Faucogney-et-la-Mer we felt that we had finally the time to explore.
As a duo we strive to connect illustration with ceramic, color with material, lines with volume and visual with contextual. Usually working in the field of vessels we decided to stay in this area of objects. As a warm-up we decided to generated drawings by ourselves and then feed the text-to-image AI with the same prompt we gave ourselves. Being able to compare our creations with the ones from the AI.
We observed and slowly got a feeling for the algorithm and how it may work.
We stayed with some images generated (by us or the algorithm) and took one step further. Coming up with more and more detailed prompts we fed into different AIs to generate designs of cups that we wanted to have as output—working hard on our skill to perform the craft of prompting. On the way we found many tools we could use, artists and designers working with AI and communities occupied by the wish to make AI more equal and just.
One step further. Giving the algorithm an image to work with. To alter, to change and modify. Picking one of the generated objects and altering it again creates a loop of creating and reacting between the algorithm and us as human designers.
Observing the current ability of the AI to generate images we were inspired by glitches and the output we deemed imperfect—making the AI a sympathetic and maybe even humorous partner to work with.
And then finally bringing something from a drawing with words to be generated by an AI and eventually molded with clay into the analog world and transforming it into something we can touch. Followed by feeding it as an image back into the digital world to be processed again. Creating a never ending loop of creation and decision making—choosing with which output we want to keep on going and creating. Lost in translation between the analog and the digital. Always wondering:
Is it possible to truly cooperate with an AI on eye level?
Reflecting on our role as decision-makers in this process. Always filtering, classifying, reordering and choosing with which output to work further—making it our next input.
The steps we took and the exercises we made within this exploration are just the first to come followed by many more. During the process we stumbled upon many questions and not all could be answered.
To finish this blogpost we like to share some of them with you—an attempt to drag you deep into the latent space in our human heads.
How can Machine Learning be ethical and just if already the dataset it is built upon reflects a white, male eurocentric world view?
What is the role of people training AI systems and what is our impact and responsibility as users?
What are the implications of copyright and ownership for a tool being able to copy styles and brushstrokes from one specific artist/designer/style?
Can text-to-image machine learning be part of our design process (Gestaltungsprozess)?
What restrictions are required to prevent people to use this technology in a harmful and/or violent way?
How can we dream beyond how AI is used?
AND ALSO
Danke an Villa Villekulla, die uns eingeladen haben an diesen wunderschönen Ort und es geschafft haben uns mit diesen inspirierenden Menschen zusammenzubringen. Merci Ali, Anita und Cheyenne fürs gemeinsame Verweilen und Gedanken spinnen.
Nikos Probst für die Einladung zum Midjourney Discord Server.
Amelie Goldfuß für den Zugang zu DALLE 2.
Die Plattform DreamingBeyondAI für den schönen und inspirierenden Launch-Event und die wundervolle Plattform.Danke an all die Menschen, die ihre Zeit in das Entwickeln von Tools stecken und Informationen aufarbeiten und teilen, um diese anderen zugänglich zu machen.
Danke an Menschen, die sich mit ethischen, sozialen, gesellschaftlichen und ökologischen Auswirkungen von Technologien beschäftigen, diese kritisch hinterfragen und nach neuen Wegen suchen.