Enjoy 2 free articles a month. For unlimited access, get a membership now.

FRAME 157 embraces AI as an imperfect, collaborative creative partner

BOOKMARK ARTICLE

Underscoring the inspiration behind FRAME 157's design, our design director Barbara Iwanicka shares learnings from early encounters with AI and makes a point for looking at the creative potential in the technology's limitations.

You are a thing, and I’m inspired by your limitations. 

There's something incredibly fascinating about the way AI creates new content – not the perfect, polished, ready-to-publish type, but output that is surprising, funny, and, by our standards, absolutely 'wrong’. Many years ago, before ChatGPT existed, I attended art academy and worked on a project based on children's stories. I asked my fellow students, who came from all over the globe, to read Little Red Riding Hood in English while I recorded it using a speech recognition tool. The tool could only recognize 'perfect English’, so when faced with strong foreign accents, it transcribed very different stories than what was originally read. The title was most often transcribed as Red Google. I fell in love with the tool’s limitations, as they turned out to be its greatest source of creativity. It behaved a bit like a small child – when they don’t know something, they usually try anyway, as there is nothing to lose – a quality we adults often see as a flaw.

This edition of FRAME deals with the presence of artificial intelligence technologies in the design world. As fascinating as AI is, it can also be difficult to grasp how it all works, and where output originates from. So I was thrilled when I stumbled upon the book You Look Like a Thing and I Love You by Janelle Shane, which is filled with often silly but very insightful and thought-provoking experiments that capture the essence of AI. Shane's approach is to test everything: AI’s potential, limitations, creativity, empathy and ability to learn. She attempts to teach AI to crack jokes, create cooking recipes and more. The title of the book even comes from a phrase generated by a neural network trying to create pickup lines. What I found most inspiring is Shane's exploration of the difference between supervised and unsupervised machine learning. There was very little AI could do before it was fed existing examples, extended sets of rules, cultural references, and so forth. Therefore, instead of fearing AI as a replacement for humans, we should view it as a partner. As she puts it, 'The sheer weirdness of AI-generated output is a reminder that AI is not magic. It's a tool, and like any tool, its usefulness depends on how we use it.'

With FRAME 157, we wanted to extend this collaborative spirit to some of the design elements. We decided to use the open-source typeface Terminal Grotesque, created by Raphaël Bastide and Jérémy Landes. Terminal Grotesque is freely downloadable, usable and modifiable by anyone. This principle extends beyond code to texts, images, machines and ideas. It promotes a sense of collaboration that we found very fitting to the theme. Additionally, in computing, 'terminal' refers to a set of output devices (such as screens) or input devices (like keyboards and mice), marking the end of a network. The pixel constraint as a module element posed challenges that sometimes sparked creative solutions. 

Fingers crossed you won't find any mistakes in this issue, but if you do, perhaps consider them not as flaws but as ‘unintentional sparks of creativity’.

Get your copy now.

Unlock more inspiration and insights with FRAME

Get 2 premium articles for free each month

Create a free account