Typography, Reconsidered – Sponsor Content

AI doesn’t know how to create words.

Allow me to rephrase: AI certainly knows many words. In fact, it may know all the words. The technology leverages statistical probabilities to predict how and why any given word might be used, and where it should land in relation to a word that precedes or succeeds it. It’s then able to stream all those words together, infinitely, without tiring.

But when it comes to bringing those words to life visually—creating the literal letterforms that construct each syllable, clause, and paragraph—so far, AI stumbles. To recreate the work that designers have been doing since the days of illuminated manuscripts and the Gutenberg Bible is no simple feat—at this stage in the evolution of AI tools, the models built on statistical probabilities struggle to recreate the simple beauty of well-designed, meticulously crafted typography.

Still, designers and artists have begun to explore what this new frontier of typographic expression may look like, tapping into an array of AI-powered tools and techniques to explore the niches of the discipline. In doing so, they’re not just creating innovative and invigorating imagery. They’re delving into the fundamentals of communication, expression, perception, and the relationship between human and machine.

Here, four designers share their work and perspectives on how these tools have reshaped the way they create and explore typography—and how the process has spurred larger questions about essential human practices primed for an evolution in the AI age.

So far, AI tools have struggled to create precise letterforms. Four designers weigh in on how that struggle is offering surprising insight into the evolution of human communication.

Barney McCann

Obsolete began as a way of exploring new letterforms and typography that could be created with AI. Using Arial as a starting point, I fed these characters into a neural network that analyzed the basic shapes of each letter, and then allowed the AI to try to find the most efficient ways to transition from one letter to the next.

The typeface is the culmination of this process, where steps normally hidden behind the pixels and processors of a computer are made visible, with endless variations as the machine looks to consistently become more efficient. The result is, I think, the closest you can get to a computer’s handwriting—but it has connected with creators in other projects because of its humanness, despite being based in technology.

In general, I don’t fear the impact of AI as a creative. I have been thinking a lot about the Fauvist’s response to photography, and focusing on how we can use AI tools to lean into and communicate the actual human experience further. People don’t resonate with purely digital outputs as much as they do with work by humans. I think my best work sits between those two worlds.

Vernacular

A. A. Trabucco-Campos and Martín Azambuja

Artificial Typography is born out of the juxtaposition of the ephemeral and the timeless. This book explores the rapidly evolving field of AI through our longstanding Latin alphabet. The book contains 26 letters re-imagined by AI. Each letterform is interpreted twice through the lens of 52 iconic artists across various media (painting, sculpture, textile). The typographic space is especially great for this exploration, since it’s a world of ideas where general conventions rule, but where there is endless opportunity for unexpected interpretations, with new recipes arising every day. At its heart, the book is driven by a curiosity to see how far AI could push visual language, and it holds many surprises, especially when the structure of letterforms are combined with materials like stone (e.g. Noguchi) or even light installations (e.g. Bruce Nauman).

Initially, the authors were also enamored by the idea of “conversation” and played with it as a title. The exchange that happens with AI machines is a form of conversation, and perhaps one of the most intellectually satisfying visual-verbal connections that has been devised between humans and machines.

Today’s AI is powerful, but it is limited. The role of the designer is far from obsolete. AI necessitates a point of origin, a thought, an idea, an editor, a curator, someone who can guide it. It also necessitates an endpoint, too—a use for the output, a reason for its imagery. This might change as it gets more powerful. At the moment, it needs the human mind more than the human needs it.

Artificial Typography is the first title by Vernacular, a small independent publisher with a focus on the intersection of form, typography and visual culture that stretch beyond the commonly agreed registers of “good design.” It was founded in 2022 by Porto Rocha senior designer Martín Azambuja with Pentagram partner + designer Andrea Trabucco-Campos. www.vernacular.is

Khyati Trehan

Creative AI tools don’t make me feel creative. The text-to-image model is efficient but takes the joy out of the making. It takes spending time on a piece, sleeping on it, making mistakes, solving them, and tons of iteration to take it to a place where it reflects me just as much as it surprises me. My workaround to settle this feeling is to deliberately make a “back and forth” between traditional tools and AI tools a part of the process, which makes for interesting workflow possibilities. When do you move from one tool to the other? What tools help you relinquish control and when is the right time to bring the work back into your hands? This ping ponging makes a tool out of a “generate” button. It’s also just really fun to play a game of exquisite corpse with GenAI instead of treating it as a means to a completed end.

What still fascinates me about generative AI is that its interface is essentially plain text. In my opinion, writers make better prompt engineers than creatives do. Working with AI tools in turn has also made me better at articulating my work and made me more intentional with the words I place on a page.

All of my type-specific outputs currently slot under the category of expressive typography rather than practical applications. In the same way new UI typesetting paradigms emerged from the rise of app-populated smartphones, I see LLM formatting rules and best practices becoming the next place for typography to serve.

See more of Khyati’s ongoing experimentation with AI tools: @khyatitrehan

Gianpaolo Tucci

Since AI image generators became a mass-market product, an emphasis on image development has seen huge progress in the algorithms that serve as the foundation for AI outputs. A year ago, the outputs were clearly a blurred and distorted version of reality—a visual dreamer. Today, realism is on the rise in AI’s mimicry of daily scenes, especially in the context of photography shoots, products, and sci-fi images.

Since I began using these technologies, my focus has been on letter shapes—typography. Typography itself is a technology for communication, but so far it has not been at the center of this AI evolution. It’s ripe for experimentation.

The whole project finds its conception in the intersection between three pillars: Time as context, nature as a metaphor, and evolution as a constant. My journey started in trying to reach “perfect” letter shapes. By perfection, I mean the value of readability and their association to what we can read.

In this process, I’ve discovered instead the beauty of what I’ve called “aesthetic imperfections”—a new opportunity for letters and their combination into words that are kissing the rigor of the function of typography while also pointing toward an enhanced visual meaning. Even if not fully defined, the letters are cognitively readable through their composition, and by an enriched visual expression.

AI offers several opportunities to explore the appearance of words and visual messages to enhance communication. AI could standardize, but could also open up creative opportunities. Our duty is to shape the control and the coexistence.

AI today is seen as an anti-human technology. But I would say AI is not a replacement—it could be a tool or an amplifier, one that may help us to deliver a more human and inclusive future, if it’s well regulated and integrated.


link