CCB Magazine

AI tools for design will soon be indispensable. Some are already used by default. How do they work and what can be done with them? And above all, how are they changing the design profession? Together with Marc Engenhart, Sebastian Löwe, until recently Professor of Design Management at the Mediadesign Hochschule in Berlin, has written a foundational work on AI and design. We talked to him about it.  

INTERVIEW  Boris Messing    

CCB Magazine: Hello Sebastian, you were Professor of Design Management at Mediadesign Hochschule in Berlin until June 2023. Last year, you wrote a book on design and artificial intelligence together with Marc Engenhart. How did that come about?   

Sebastian Löwe: I started working as a research assistant at HMKW in Berlin in 2017 and came across this topic through a conference. At that time, only a few use cases were discussed, there was no overall view of the topic yet. Two years later, I joined forces with my colleague Marc Engenhart and we decided to write a foundational book on AI and design. Our book is the first of its kind that provides information about the entire design process and, in particular, also addresses the question of how to actually design for AI. That is, when you look at AI as design material in itself. We wrote the book on the one hand to make the topic clear to ourselves. On the other hand, to give others a systematic overview of the topic. The book is aimed at both academics and designers. We also published the website designundki.de, where all AI tools are listed and updated.

CCB Magazine:Reading the book, I realized how advanced AI tools already are in the design field and how rapidly they are evolving. But how widespread are they in practice already? Are AI tools being used by design agencies as standard? 

Sebastian Löwe:We did a small user survey in the context of the book and found that so far only a few designers use AI tools. That was about a year and a half ago. Today, there are probably many more, because the development is going very fast. With Fusion 360 or Spacemaker from Autodesk, for example, there is an AI that has become standard. These tools are aimed at product designers. With their help, someone building a chair can see what physical forces are acting on it, under what circumstances it will break, for example. The design of the chair can be adapted accordingly. Natural forces such as wind or noise, which you as a designer cannot see when you design an object, are calculated by the AI. They can also calculate whether there is enough light in a courtyard for trees to grow, what the heat development of a house façade is like, or how exhaust fumes develop in a room, and so on. All of this is already available on the market as standard.

AI technology does two principal things: automation and augmentation. Automation allows designers to focus on other things and saves time; augmentation strengthens their expertise

CCB Magazine:The book lists a whole range of AI tools. I was particularly impressed by tools that can transform 2D images into three-dimensional spaces. Or AIs that create images from text descriptions. So what does this mean for the design field? How are these tools integrated into the design process?  

Sebastian Löwe:When we wrote the book, none of this was a big thing. OpenAI just had the DALL-E project in the pipeline, an AI tool that, as you just said, generates images from text descriptions. Today, these models are available everywhere: these so-called diffusion models create images from a cloud of pixels. It's combined with a text model, and it's only in the interaction of text-based generative AI and these diffusion models that you get what we know with Midjourney, Stability.AI, and DALL-E 2. Honestly, these tools aren't really ready yet. But the exciting thing is that you can train them yourself, provided you have some budget as an organization. Many agencies are starting to train their own AI tools to get the results they want. As an individual designer, of course, this is not possible. Nevertheless, independent designers can use many useful tools for their work.

CCB Magazine:How does this change the way designers work? Does this also change their understanding of creativity and design?

Sebastian Löwe:What we see is that technology is doing two things. One is automation. We know this from other areas of the economy. This mostly concerns things that are not exciting for the designers to do anyway, such as cropping in image processing or various other preliminary work. Automation allows designers to focus on other things and saves time. There are also tools that can help with research and organize a huge amount of data in a meaningful way. The second dimension is what we call augmentation. Augmentation is where the expertise of designers is amplified. This is a very exciting area! Even experienced designers reach cognitive limits. Earlier I mentioned forces of nature that can affect design. It's hard for a designer to calculate something like that. That's the crazy thing about this technology: any statistical relationship can be mapped and learned and then generalized. This means that a forecast can be created from all possible data. This makes it possible to predict, for example, whether people will recognize brands in a particular design. Through tracking, the AI determines what users like and what they don't like. There are studies on engine hoods, for example, where AI can map exactly which part of the hood is perceived by the customer as a trademark. As a designer, you can't understand that at all because it exceeds your cognitive possibilities. 

Through AI tools, any statistical relationship can be mapped, learned, and subsequently generalized. This means that a prediction can be made from all possible data. This makes it possible, for example, to predict whether people will recognize brands or what users like and dislike about a particular design

CCB Magazine:It can be said with certainty that sooner or later no designer will be able to avoid AI. But are there any further training opportunities or even design degree courses that provide a practical introduction to the topic? 

Sebastian Löwe:This is being hotly debated in the academic world right now. For example, there is a research project called KITeGG, which involves five or six design universities that are considering how design education must change as a result of AI technology. Does it have to change at all and in what form? They are trying to familiarize the students with the tools as quickly as possible. My colleague Marc and I have already taught the subject at our universities. But many are also following the development on their own. There are numerous online continuing education courses that offer prompt engineering, for example, or explain how to use generative AI tools. If you do some research, you will find what you are looking for.

CCB Magazine:To come back to the automation processes. Aren't valuable skills possibly lost through these AI tools? As a designer, aren't you giving up too much autonomy if you always rely on the results of the AI?

Sebastian Löwe:You still have to have the expertise as a designer. These AI tools make thousands of suggestions that need to be evaluated and curated. And they make mistakes. So you have to decide which ideas to pick up and follow up on. This kind of curatorial intervention requires a high level of expertise. Knowing what works, what has design value, what is ethical, the aesthetic design basics, in other words, those still have to be learned. And one way to access these basics is through drawing. To create a connection between hand, eye and mind - that is the basis for a professional designer.

The most exciting developments are currently taking place in EmotionAI research, where the aim is to recognize how someone is feeling emotionally in order to adjust the design in real time

CCB Magazine:How does AI change the experience of digital products? What opportunities does AI open up for users? Keyword: personalized or individualized design.

Sebastian Löwe:Personalized design works in principle with anything that is based on data and can be digitized. In the case of websites, the design can even be customized in real time. Depending on your wishes and preferences and my wishes and preferences, the design of the website then changes. That's a whole new user experience! The most exciting developments are currently taking place in EmotionAI research, where the aim is to recognize how someone is feeling emotionally and then adjust the design in real time.

CCB Magazine:Is it rather big players that develop the AI systems like Adobe and Autodesk or is it many small developers? What new business models will result from this?

Sebastian Löwe:Both. There are many strategies. Meta, for example, does everything via open source and wants the developers to be on board. The developers benefit from the models that Meta provides, and Meta benefits from the wealth of ideas that this generates. Later, they can charge the developers a certain percentage of the tools used by the users. Or you do it like Midjourney and say you pay twenty euros a month or so to be allowed to use the software. There are different business models. Everyone has their own strategy to create market penetration.

CCB Magazine:Finally, a prediction please: On balance, will AI in design create more or fewer jobs? Which jobs will be eliminated? Which new ones will be created?

Sebastian Löwe:I prefer to leave the speculation about jobs to the studies of large consulting firms. I have absolutely no understanding of how that will develop. Certainly, it is always a hardship of new technologies that they make the work of some people superfluous. What we can already foresee is that the executive functions of designers are becoming less and less. The work is shifting to a strategic, curatorial, managerial level. And the other thing is that as a designer you have to learn how to use these AI tools.