In 2023 I designed and developped the website for the research project Design & Machine Learning, using then-cutting edge generative AI.

Design & Machine Learning was a research project which examined the mutations of graphic design in contact with contemporary artificial intelligence.

The goals were (I) to grasp the current uses (projects, opportunities, etc.) and problems (blockages, shortcomings, etc.) of machine learning technologies for designers, and (II) to initiate a “research-creation” process to set up prospective scenarios adapted to independent design studios.

Several outputs were produced:
– 15 interviews with European graphic design studios pioneering the use of machine learning, conducted by Alexia Mathieu
– A multi-media theoretical essay: Artificial Design: Creation Versus Machine Learning
– A ChatGPT-augmented website design-machine-learning.ch
– An innovative communication campaign created with AI (“Davinci’s insights”): a poster series and 3D short film, awarded at the Biennale de design graphique de Chaumont (France)

The essay Artificial Design: Creation Versus Machine Learning by Anthony Masure explores three research questions: How can the paradigm of automation be redirected on a smaller scale, in a “tailor-made” way? How can machine learning foster invention and curiosity? How can design help to counteract the dominant culture of AI? The essay is available in various open access digital formats: PDF, HTML, ePub, MP3.

“To ask whether artificial intelligences can create is not framing the problem correctly, firstly because AIs are in no way magical, and secondly, because the vocabulary of creation, when detached from material contingencies, gives rise to an incapacity to envision technique as a space of exploration, and thus, of ‘production’.” – Anthony Masure, Artificial Design: Creation Versus Machine Learning, p. 102

“One must deconstruct two other myths regarding so-called creative artificial intelligences: the first being that artists or designers have total control over their production, and the second that the machine can be totally autonomous. A more interesting way to proceed would be to explore the scope and location of chance when introduced into the phases of production.” –Anthony Masure, Artificial Design: Creation Versus Machine Learning, p. 80

The website design-machine-learning.ch contains 15 interviews by Alexia Mathieu with artists and graphic design studios pioneering the use of machine learning: Boyd Rotgans, Cristobal Valenzuela, Deniz Kurt, Feileacan McCormick, Kévin Donnot, Marta Revuelta, Martin Tricaud, Meredith Thomas, Nadia Piet, Nicolas Barradeau, Rifke Sadleir, Simone Rebaudengo, Superposition Studio, Étienne Mineur. The goal was to better understand the current state of the art and the problems associated with integrating artificial intelligence into a graphic design workflow.

Artist and designer Douglas Edric Stanley created an experimental website (design-machine-learning.ch) that takes all the texts (interviews + essays) from the research project and remixes them in real time using text and image generators. The interface was inspired by Smalltalk-76, the first interface of a complete desktop publishing system.

“Something that will continue to happen is that designers will have more of a role in curating: deciding where to go but also what to feed these algorithms. Depending on what you feed, you’ll get different answers. You’re also curating your own algorithms. What you put into the system to train will influence what you get out of it. And so, you’re an editor at the beginning and you’re an editor at the end.” – Cristobal Valenzuela, Runway

“If you want to do something more advanced, you’re going to have to delve a lot deeper into the tools and this can be technically quite demanding. These tools aren’t meant to replace us, they are an expansion of the toolkit, the pencil is still relevant. There are things you can do with a pencil that you cannot do easily with a neural network, and never will.” – Feileacan McCormick, Artist

“It really surprises me sometimes when I do things that are supposed to be really wrong. I don’t use well cleaned data sets on purpose. Ev- eryone tells you that you must have a clean data set because that’s the way to get good outputs. What is interesting for me is not good outputs but exploring even the data set as a medium to express new ideas”. – Simone Rebaudengo, Designer

“The way people talk about the models they’ve trained, it’s like people talking about their babies. If you raise a child and you’re really horrible to it, it will probably turn into an adult who sees the world as this horrible place and acts accordingly. And it’s kind of the same with machine learning — what you input into it is kind of chaotic. And before you fully train it up, it will say some really ropey, out there, stuff. Kind of like a toddler or something.” – Rifke Sadleir, Designer

“What’s interesting is that, you can have a conversation with anything, you can have a conversation with Pluto as a planet, and suddenly you have a bot basically impersonating and having a point of view of a planet.” – Simone Rebaudengo, Designer

Graphic designers Élise Gay and Kévin Donnot (E+K) were invited to produce a carte blanche communication campaign for the essay Artificial Design. Their proposal, entitled Davinci’s Insights, questions the relationship between graphic design and AI through a series of five posters and a 3D short film.

E+K: We want to explore the possible images that could convey the links between design and machine learning. How would you picture these issues and their interplay?
Davinci: With a Rubik’s Cube
E+K: Why?
Davinci: This object symbolizes the complexity of machine learning and its ability to solve complex problems.
E+K: That’s a game. Why do you think that a game is relevant to depict machine learning and design?
Davinci: A Rubik’s Cube is a great object to represent the complexity of machine learning because it is a game that requires problem-solving skills and analytical thinking. Just like machine learning, solving a Rubik’s Cube requires an understanding of algorithms.
E+K, Davinci’s Insights