+
Generic selectors
Exact matches only
Search in title
Search in content
Post Type Selectors
Who's In This Podcast
Helen Todd is co-founder and CEO of Sociality Squared and the human behind Creativity Squared.
Martin and Moritz are co-founders are Process Studio, a design studio based in Vienna, specializing in generative and interactive design.

Ep32. Martin Groedl and Moritz Resl: Are Machines Creative?

Up Next
Ep33. Nem Perez, Sway Molina & Ben Nash: "OUR T2 REMAKE"

Ep32. Are Machines Creative? Discover the Evolving Answer Due to GenAI with Process Studio Co-founders Martin Groedl and Moritz Resl

Martin Grödl and Moritz Resl founded Process Studio, an experimental design studio based in Vienna that specializes in generative and interactive design for branding, web, installation, and print. As well as traditional graphic design solutions, Process designs and develops highly specialized software that is used as tools for and by clients who include MIT, Design Museum Holon, MAK – Museum of Applied Arts Vienna, The Royal Museums of Fine Arts of Belgium, and more. 

In 2021, Process designed the official Austrian contribution to the London Design Biennale. In 2022 Grödl and Resl were appointed as guest professors in visual communication at FH Salzburg. Their work has been featured in design publications like Dezeen, WEAVE magazine, The Type Directors Club Annual, Gizmodo, and The New York Times among others.

Martin is a designer, programmer, and artist who works on the boundaries between Design, Art, and Digital Technology. He participated multiple times in Google’s Summer of Code working on further development of the Processing programming language and was a guest artist and researcher for Motion Bank in Frankfurt. He taught courses including Introduction to Programming, Physical Computing with Arduino and Generative Graphic Design for Printmaking at the University of Applied Arts Vienna and tutored courses including Data-modeling, Database systems, Distributed Systems, and Semi-structured Data at the Vienna University of Technology. 

Martin Grödl holds a Master’s degree from the University of Applied Arts Vienna and a Bachelor’s degree from Vienna University of Technology.

Moritz is a Creative Director and Designer who Works at the intersection of Design, Art and Digital Technology. Previous to co-founding Process, Moritz worked for Sagmeister & Walsh in New York City, where he designed for clients like Adobe and The Jewish Museum in Manhattan. 

Moritz holds Master’s degrees in Computer Science and Media Art from the Vienna University of Technology and from the University of Applied Arts Vienna. His work has won international awards including Type Directors Club, ARC, and LACP Vision awards. 

When they finished their degrees, they realized the lack of career opportunities in their shared passion, and so, they forged their own way, focusing on designing and developing highly specialized software tools for their clients.

“We had to make our own tools in order to realize what we thought of in the first place for our projects.”

Moritz Resl

They marry their technical backgrounds in statistics, mathematics, and programming with their eye for design to carve out a unique place in digital art and computer science. 

Pioneering their Way Into A.I.

GenAI’s explosion this last year put it on everyone’s maps. Martin and Moritz were years ahead of the boom, not only dabbling, but developing artificial intelligence tools years before anyone had ever heard of ChatGPT.

One of their most prominent projects, UNCANNY VALUES: Artificial Intelligence & You, was featured as part of the Vienna Biennale for Change 2019. MAK — Museum of Applied Arts Vienna, commissioned the team to produce a key visual identity for their exhibition on artificial intelligence. Around the same time there was a project started by Nvidia, where they were recreating hyper-realistic human faces instantly. To achieve this, the team at Nvidia used top-of-the-line hardware and expensive equipment to achieve their results. Martin and Moritz only had their individual laptops. 

The primary challenge? Finding a way of using less data without access to high-performance equipment but getting a similar result. The solution? emojis.

The duo set out to train a neural network that would generate low-resolution emojis, like distorted versions of the little yellow characters we know and love. They call them AImojis

They built a dataset of diverse emojis scraped from web pages and mixed the statistical patterns to teach a model to create new visual output. Moritz describes the result as “very weird little faces.” 

A high-quality, streamlined result was not the goal. Instead, the AImojis look blurry, confusing, and grungy. Martin and Moritz celebrated this low-res approach and showcased it over the city on large banners and posters, blowing them up enormously, to highlight the poor resolution and create shock value beyond emojis’ usual domain within our screens.

AImojis

Grungy, unrefined, weird, uncanny, and broken are some of the words they use to describe the AImojis. Created on earlier laptops, Martin fondly describes the low-res characters as being created by “kindergarten A.I.”  

“We like that approach of what comes out if you remove all the polish, and all the guidance, and all the filtering, to what is really at the core. It’s dirty, it’s messy, it’s grungy, but we liked it.”

Martin Grödl

UNCANNY VALUES contemplates the emotional exchange between human and tech. A.I. is undoubtedly becoming more capable every day, but for what purpose, if not humanity’s?

“I really think this symbolizes our relationship to technology in a very interesting way. Because we are not separate from technology, we are the ones who produce these A.I.’s, and we choose to make it happy and confused and angry and greedy. So this is a good reflection, all these emotions, these emojis, thrown at us, is a reflection of ourselves.”

Martin Grödl

Martin and Moritz lean in favor of A.I.’s dependability on the power of humans. Even if an emoji has a clear representation, we as humans still have the power to interpret it as we please, from person to person. Different cultures, generations, friend groups, and families may all use the same emojis in different ways.

When it comes to these uncanny, slightly disturbing AImojis, the added layers of not being able to quite understand what emotion they’re trying to convey is probably more reflective of human emotion than the emojis that we use every day in our phones that have these (mostly) assigned values.

So, not only are these funky AImojis challenging our normal dialogue, and expanding it with new options, and yet they’re not easily put into a box. At the end of the day, AImojis, among the countless other forms of communication and language, are all just different types of expression, powered and controlled by humans.

AIfont

With the system successfully creating these evocative visual emojis, the duo decided that their next foray into building A.I. tools would be in the world of typefaces. They used images containing texts of various fonts to generate new typefaces, which resulted in an endlessly oscillating visual stream of new fonts. In the process, they created one that captured the spirit of their exhibition, dubbing it AIfont.

AIfont was born quite naturally and was technically similar to AImojis. Already having the system set up, they started testing typefaces instead of emoji faces. Especially for someone in a creative field, unlocking a tool to create a new font is valuable to have in an arsenal.

“Of course, type is such an important vehicle for transporting information.”

Moritz Resl

Currently, anyone working with A.I. tools knows that they can produce images with incredible depth and detail, but prominent image generation models like Midjourney still struggle to generate images with legible text. Meanwhile, Martin and Moritz’s model was reliably producing stylized text four years ago, long before the recent GenAI boom. 

“We basically tricked the system.”

Martin Grödl

Essentially, they hijacked a GenAI system, which is not specifically focused on fonts, and once again trained it on a dataset of individual font characters scraped from the internet. This created a font, but with images of words or a sentence instead of just text.

The team was getting requests to use the font, but these were roadblocked by the cumbersome process of sending text, generating images, and sending it back. And so they decided, just this year, to make a real font out of that, which is AIfont, a usable font that anybody can download, install, and use in normal design and text processing programs. A typical typeface family consists of around 50 fonts, whereas AIfont has 500. Not all of them are finished fonts, so they call it a “typographic experiment,” offering a variety of bold new designs rather than professional tools. Martin shared that their approach was much more from an experimental, technical, logical, generative point of view instead of producing a perfect font.   

“This is still more an A.I. project than a font project. It again shows this kind of low-level baby A.I. without supervision. Without reinforcement, without selection, it’s just reproducing what we put in, and in this case, a lot of fonts, a lot of rendered glyphs. And then it tries to reproduce.”

Martin Grödl

Utilizing ‘Traditional’ A.I. Tools

MidJourney, ChatGPT, Dall-E, and dozens of other A.I. tools at our disposal have given everyone, even non-technical users, the opportunity to play in the artificial intelligence sandbox. Martin and Moritz prefer to work on and use their own models. 

Moritz explains that, while they definitely dabble to stay up-to-date and in the know, that’s about as far as they go with widely-used A.I. tools. 

“We are playing a little bit. But only from a very curious standpoint. We’ve tried them out, but it got quite stale quite fast.”

Moritz Resl

Moritz says the decision about what tool to use should depend on the project, acknowledging that tools like ChatGPT and Midjourney are powerful creative aids. Martin says he sees the transformative power of advanced GenAI accessible to everyone. 

“I’m mostly interested in how these tools will change society, jobs, and things like that, which is hard to anticipate, of course, but it also seems obvious that it will have a big impact, right?”

Martin Grödl

Designers First

Despite their degrees, experience, and expertise in the art community, Martin and Moritz say they don’t consider themselves as A.I. artists. They identify as designers by craft, who take new technological developments and incorporate them into their workflow. 

“Mostly we take a step back and really try to analyze what it is that this technology makes special. And then we try to recreate it ourselves and make it somewhat our own.”

Moritz Resl

Their passions align with simple, bold, low-resolution designs rather than high fidelity, seemingly ‘perfect’ results that mainstream GenAI tools produce.

University Students Ask: Will A.I. Erase Design Jobs?

Generative design is sneaking its way slowly but steadily into the design curricula everywhere, which is not surprise considering all that’s happening with generative A.I. that school systems would be adopting it much faster now.

So, when faced with the questions of ‘how is my future job affected by this industry?’ by current design students, they shared their opinions.

Either you embrace it fully, dive into learning these tools to get ahead of the curve and focus primarily on the skill set of prompt writing or data curating, which they foresee as being a prominent job in the future. 

Another perspective worth considering, they predict, would be to take a step back and think of one’s role as a designer more as a directorial role instead. As a curation role of defining a concept, rather than producing a specific image or a specific text. 

“With all these tools at our fingertips, mass-creating low-effort projects will only get easier, but as soon as you’re talking about interesting design challenges, I think it’s so important to have a curatorial role. Somebody who talks to a client, and understands them, and translates what they want.”

Martin Grödl

Martin also says that younger generations have an inherent advantage of being able to adapt to new technology as it arises. 

Can Machines Be Creative?

Before the rise of these tools, Martin’s answer to this question came easy – no. 

“Before, let’s say, ChatGPT, when I was asked this question, I always said, no, these machines are not creative. Not at all. We build a system, we as humans, or a society, or as a company, we build a system that produces stuff. So in the end, it’s us, who are the creatives, and this is just a mechanical, deterministic tool.”

Martin Grödl

But he says his perspective has shifted. ChatGPT is gradually becoming a general-purpose assistant with computer vision and listening capabilities. LLMs and other models are flawless at reproducing, much like the creative geniuses of our past, and their predecessors. They all interacted, learned from, and built upon the knowledge that was there before them.

Ultimately, Martin says that machines are creative in some sense; their limitless power to produce mass quantities of images, for example, would take a human an incomparable amount of time longer to produce. Past that one element of creativity, however, Martin argues that the human is the creative force.

“That’s only one thing, that’s only one part of creativity, all that personal style, all that personal experience, all that values that you learn, this comes into play, as well.”

Martin Grödl

It can’t be forgotten that these tools, regardless of their level of creativity, are mirrors of who we are individually and as a society and that we’ve all kind of got to decide together what that means. Even if things are generated by A.I., people still need to use their own personal judgment and critical thinking skills to move through life.

“We always have to think, what kind of future do we want? And what is my work, my API, my tool, my website, or my project, what kind of future is it promoting?”

Martin Grödl

Links Mentioned in this Podcast

Continue the Conversation

Thank you, Martin and Moritz, for being our guest on Creativity Squared. 

This show is produced and made possible by the team at PLAY Audio Agency: https://playaudioagency.com.  

Creativity Squared is brought to you by Sociality Squared, a social media agency who understands the magic of bringing people together around what they value and love: http://socialitysquared.com.

Because it’s important to support artists, 10% of all revenue Creativity Squared generates will go to ArtsWave, a nationally recognized non-profit that supports over 150 arts organizations, projects, and independent artists.

Join Creativity Squared’s free weekly newsletter and become a premium supporter here.