How AI might impact Design

Jonathan:

Welcome to CCTalks, the Cleverclip Podcast. I'm your host, Jonathan Tilly. And today we've got an exciting episode for you. Now, this is our premier episode of CC Talks - the Cleverclip Podcast. So you might be wondering who we are and what we do. Well, here at Cleverclip, we create animated videos and interactive experiences that help explain a complex topic and inspire your audience with an idea. And we're also doing that in an audio format with our podcasts, CCTalks - the Cleverclip Podcast. So today on the podcast, we want to explain a complex topic and help you wrap your head around artificial intelligence or “AI” for short. And what better way to do that than with not one but two special guests who will dive into how artificial intelligence impacts design. Both are from Statworx, professional data science and AI service for businesses. Livia Eichenberger and Stephan Müller are on the cutting edge of all things AI.  Livia, and Stephan, welcome to the show.

Livia:

Hi, thank you.

Stephan:

Hi Jonathan. Thanks. Great to be here.

Jonathan:

It's so good to have you guys here. So, um, I'm sure I could go into all the amazing things that you guys have done, but, um, would you like to tell us a little bit about yourself?

Livia:

Yeah, well, of course, uh, shall I start?

Jonathan:

Sure, great.

Livia:

Okay. So hi, thanks for having us here. As Jonathan introduced me, my name is Livia. I'm working at Statworx. I'm a business development and innovation specialist there. Before that, I worked as a consultant data science, where I worked on various data science and artificial intelligence projects. Uh, at the moment we're building up our location in Zurich. So together with Stefan and my other colleagues, we're trying to, yeah, make our office bigger there. In my role, I’m also the team leader or, of our new initiative, uh, AI and society. Uh, with this initiative, we have the goal to educate the broad population about AI. We want to analyze and address, address the impacts that AI has on the whole society. So I think it's really great and I'm really glad that we can be here to address the population and talk about this topic and how AI impacts design. Um, it's also my private interest because in my free time I've been working as a professional photographer for almost 10 years. So I kind of, we have two kind of overlap of two of my biggest interests. So I'm really glad to be here and look forward to this discussion. Thank you.

Jonathan:

Oh, fantastic. So great to have you on the podcast. Stephan, you're up next?

Stephan:

Yeah, great Livia, it was really interesting to hear also about the AI society. It’s a great thing that's going on at our company now. I also, I work as a consultant data science at Statworx, ehm, still in our core business where we solve business problems using AI and machine learning technology. Especially, I focus on deep learning, which is a subfield of artificial intelligence and the fields that I'm most passionate about are natural language. So how can a machine learn to understand language, to generate language? How can we interact with this kind of machine and also forecasting. So how can we try and see into the future with machine learning algorithms? Exactly. Yeah. I'm a bit more about my background. I actually started off my studies or my academic, um, academic career with, uh, studying electrical engineering. And then in my third year I worked a lot on robotics and then I also, uh, then I found out more about the field of data science and just, it's like, it's endless variety what you can use it for that and fascinated me. So I decided to pursue my master's degree in data science and yeah, exactly. And there, I also focused on gangs, which are a kind of network where you can generate new data, which are also used a lot in design. So I'm really excited to talk more about that today. Yeah. Maybe private, uh, I'm very interested in sports and I do a lot of sports myself. I play field hockey and I'm very passionate about endurance sports as well.

Jonathan:

Wow, fantastic. So great to have you both on the show, Livia and Stephan. Um, I know for, for me artificial intelligence, you, you just hear it and it's such a broad topic and you know, most of us we've already encountered AI, whether it be asking Alexa, play my favorite playlist or a chat bot in customer service, right? We all understand what that is, but there's so much more to AI than just that. And that's why as a creative agency and experts in the field, design, we at Cleverclip, we want to dive deeper into the design field and find the link between AI and design and ask questions like, uh, will AI influence design process in the future, uh, what opportunities are going to arise and what challenges are going to be expected. But before we dive all into that, right, let's get the general foundation about AI and how it links to, to design. So artificial intelligence, it's developed into a buzzword and it's a frequently debated topic in different areas in different industries. And most of us may think about it as a somewhat science fiction based robot concept to seen in movies. So to start off, how would you two explain and describe artificial intelligence to a normal person out there?

Livia:

Okay. Yeah. Just to say a word about those movies. So these kinds of robots taking over the world, we're not there yet, so maybe we'll go there someday, but we're not there yet, but yeah, you said it correctly. AI is a buzzword and actually no one really knows, I guess what it exactly is, or they're kind of different myths. Um, but let me start from the beginning. So AI stands for artificial intelligence, so artificial, meaning non-natural, non-human intelligence so artificially created intelligence. So in practice, those are systems computers that try to mimic the human intelligence. So as Stephan already said, like for example, understanding, speaking, communicating, reasoning, uh, recognizing things. So all those kinds of things. However, um, it's also difficult to find one definition for artificial intelligence because even human intelligence is not clearly defined. So what is human intelligence is? Is it just like, if I have an IQ of, I don't know, 150 or higher, is that intelligence or, um, is, does it also have to do with emotions? Is there something like emotional intelligence? So as long as we don't have this definition for human intelligence, it's difficult to even start to pin down what artificial intelligence exactly is. So, um, yeah, I think it will stay kind of a buzzword, but I hope to have explained a bit what kind of what's the aim of those systems. So really to try to mimic human behavior, human intelligence to solve, uh, and yeah, problems and tasks.

Jonathan:

Yeah. Stephan?

Stephan:

Yeah. So you can actually take it a bit further from there in the, and artificial intelligence, we have different ways of seeing, like seeing artificial intelligence by the complexity. So we like can start with weak AI where we say, okay, we focus on one simple problem. Like for example, understanding language, being able to make sense of, uh, sentences or whole texts and yeah, exactly. So we focus on a specific problem and how that can be solved. So that's like a part of this, uh, artificial intelligence, but then there is a general artificial intelligence where we try and go further and say, how can we act like a human? So not only solve a specific problem, but actually like more in general, like act like a human, maybe reason, like a human make decisions without having been explicitly told what the problem is. Exactly. Yeah. And then there's actually also a concept called “super AI” where we say, okay, what happens if an AI can do even more than we humans can do that opens a whole another variety of topics. Possibilities. Yeah.

Livia:

Yeah. And it's important to say that we, we still are at the very basic level. So nowadays, we act into field of weak AI. So we're not even at general AI level. So we don't have, uh, any system that is as intelligent or let's say as kind of has this human intelligence,  so to say. So our algorithms, our AI's nowadays are really trained and built to solve one specific problem, as Stephan said, and you cannot use one AI to solve another problem. So it's really this, uh, specialists that we develop in that field, but not like a broad, um, system of intelligence. So to say.

Jonathan:

Exactly.

Stephan:

Yeah, in this field of weak AI, we can actually do certain tasks even better than a human can do or much faster. So we can even exceed human performance. But yeah, general AI is still a problem that is, um, yeah. It's even not, well, it's being researched a lot, but where we haven't made significant progress as in weak AI like solving specific problems.

Jonathan: Yeah. Now you are in Frankfurt, Germany. Correct? And Livia, you're helping build the, the expand to, to Switzerland, right?

Livia:

Yeah, exactly.

Jonathan:

I also live in, in Stuttgart Germany. So even, even the thing of, um, AI artificial intelligence in Germany, in, in the German language, it's, “künst, künstlerische Intelligenz?”

Livia:

Künstliche Intelligenz.

Jonathan:

And “K” as “Kunst”, art. So even KI and AI, I see German friends, like what's the difference between KI and AI? Is it the same thing? It's the same thing, right?

Livia:

Yes it’s the same thing. It's just like the English and German, like yeah, word for each other.

Jonathan:

Yeah, exactly. Okay, good. Just wanted to [inaudible] that.

So thank you so much for giving us the layman's terms of what AI or KI in Germany, is. Um, so AI it's it's, it is quite a broad topic and it does encompass many different disciplines that can be used in different industries. So in order to narrow it down, we want to focus this podcast on the area of design and how AI can impact the design process in the future. So let's start with some general things on how we can imagine AI being involved in the design process. So can we imagine a robot taking a pencil and drawing something on a piece of paper? Where does artificial intelligence and design meet?

Stephan:

For me, it's really important to, um, what do you mean? Do you mean if the robot is, uh, can draw, if you tell it what to draw or if the robot is able to sit down and just start drawing for itself?

Jonathan:

That’s the question.

[inaudible] Stephan:

So it's about the act of like being creative more than to be physically able to draw like yeah. Okay.

Livia:

Yeah. I mean, well maybe we can start off where we see it nowadays. Um, so you already have it in so many fields of design, not only the drawing, like the graphic design, but also interior design, industrial design. So every part or every sub field of design is already affected by AI. And as we see it today, it's like, um, AI influences, uh, this area of design in two different kind of ways. So one way is that processes can be automized. So routine work, repetitive work can be done much, much faster, and designer can concentrate more on the creative part, let's say, uh, like, uh, gathering ideas, brainstormings, drawing sketches, and then the implementation is done by the AI. So that's one part and the other one, uh, it's that AI can be seen as an artificial muse. I like this expression so that it kind of inspires the designer. Um, so it gives him impulses, new creative impulses. And there, we also go into a direction where Stephan is an expert about [inaudible]. We have this, um, algorithms that can create based on inputs, different data, then they can make new things. So kind of new pictures or new designs. So a designer can use it as an inspiration if you kind of get stuck or, uh, yeah. Just want to brainstorm or gather some prototypes. He can use those tools to let him inspire him and, uh, yeah. Use it for his process. So that's kind of more an interactive way how AI and designers could work together. So, we see dose two kind of fields where we are now where AI and design meet . And I mean, there are many different examples. Uh, so first, if we look at this kind of productive part where AI supports like the production of, uh, designs we have, for example, um, the Wix website, you know, where you can do your own, uh, website, um, until now I think, or until recently you had to build it yourself. I mean, it has been made easier for you, so you don't have to program any uh everything's it's just like you have these building blocks, but now they go even that far, that you can kind of tell them, uh, how you want your webpage to look like what's your content, what you want to upload there. And then this algorithm, uh, gives you certain layouts, uh, for you personally designed, and then you can choose from, so you don't even have to go into the process of, uh, doing everything yourself.

Jonathan:

So, okay. That's, that's a great concept for me to, to wrap my head around because you go, what, what is it and how does it help? But, but like you said, you know, just making it more intuitive and going, okay, this is what I want from a website. Okay. Here's all the different layers that we can offer you in one place.

Livia:

Yeah. And it's not even what we can offer you, but it already tries to find the best ones. So they, it won't offer you like all the ones. So if you say like, okay, I'm a photographer. I will have many pictures on my website. Uh, I have, yeah, I need high resolution. Uh, I don't need much text. Uh, it takes this into account and only shows you like, what's best for you, what suits your needs. So it also lowers the, the barrier to entry into these fields. So you don't need this kind of, to be a design expert to make your website. So it's, uh, yeah. It's, it brings design closer to everyone, I guess. Yeah.

Jonathan:

Wow. Fabulous.

Livia:

I mean, speaking of, uh, photography that's, I mean, that's where I'm coming from there. You have it, uh, in many applications as well. So, um, in Photoshop, I guess if you worked with that, um, sometimes you'd like to select part of a picture, so maybe a person in front of a landscape or something. So until now you had two options either you kind of selected the same colors somehow, so that you could extract this kind of part or that you kind of wanted to draw the line yourself, but that was very cumbersome. So now they have this tool that you concern, um,  say, “find figure”, and then it automatically perfectly selects you in, from the landscape and you can kind of cut it out or do something with a selection. So that's like, that's amazing.

Jonathan:

Yeah. So that is AI, then.

Livia:

That is AI, yes.

Jonathan:

So we're already using AI, like without us really knowing or knowing it, but again, oh, this is just a cool feature, but actually this is, this is AI in, in, in progress. Livia:

Um, and I mean, another thing we're, I think almost everyone uses every day are the Instagram filters. So the ones, the ones where something happens with your face, I think there's the one with your heart thingies on the face. I don't know. So the, the algorithms for stuff to detect your face to place the things, uh, on the correct spot. So that's also AI, that's all about the face recognition and…

Jonathan:

So, AI is everywhere then.

Livia:

Yes. Yeah, yeah.

Jonathan:

We just need to pinpoint it.

Stephan:

What's actually interesting to get back to your original question. If an AI can just sit down and start drawing, we're actually really good at is if we tell the AI a specific problem, we want solved, for example, you have an image and you say, Oh, this landscape looks beautiful. I wonder what it would look like if Picasso had drawn this, so we can teach an AI to transform this input image to an image of Picasso.

Jonathan:

Wow.

Stephan:

So that's in that style. Exactly. But what we can't really do is just let it set an AI free and let it do what it, let it do its job because we really still we're in this week phase where we have to still exactly describe what we want and the algorithm can only be as good as we can measure its performance.

Jonathan:

So it sort of sounds like AI, at least right now, and weak AI is more like a productivity tool instead of like here, grab a pencil robot and make it it's, it's more of a tool that helps us.

Livia:

Yeah. Yeah.

Jonathan:

Got it. Okay.

All right. So you, AI can, it can definitely increase productivity as, as we were talking about, and especially with the nerve and time consuming routine tasks that just simply have to be done. Right? And it can be fun as well with the Instagram filters, like, Oh, wow, that’s like this on the cheeks or hearts on my forehead.

Livia:

Yeah, yeah.

Jonathan:

So what could be the different advantages of implementing AI in the design process? So how can, how can designers benefit from AI?

Stephan:

What I find really interesting is like the concept of generative design. That you actually don't sit down and start designing something, but you like, try to give the AI a frame of what your constraints are. If we go like to more industrial design and you have to design a part, say for yeah, for a robot, then you say, okay, I have this much space and I have these forces acting on this part. And then you tell the AI, okay, these are our constraints. How would you design this part? And then the AI gives you recommendations, or it gives you just like, uh, generates ideas. And then you can take it further from there. So you don't have to start the design process from zero, but you get like a recommendations and then you can see, yeah, I actually like this. I like this, but I don't like this, but then the AI can iteratively, like generate something. That's more like to your liking and at the end, we're not there yet that these, uh, these systems work perfectly. So this, the designer will always have to do like the last, like quite good, like fix it to perfection. Yeah. Yeah. Yeah.

Livia:

And I mean, it also gives us, I think, new possibilities or things that were very cumbersome if, if you don't do it on the computer. So I think there's this shift from a bit, let's say the physical world to the computer world so that you do more things in post-production for example. So, I mean, it's not directly AI, but um, no, not AI, sorry, not design in that sense of graphic design, but, uh, when you talk about the filming industry, uh, it's very interesting. I just read it, I think few days ago, um, uh, in the past, if you wanted to age an actor or make him younger, you had to do like lots of makeup and stuff. And it was very cumbersome and took so much time. And now you have these algorithms that do that on the computer. So I think everyone, or most of the people know this app where you could age yourself, where you could see how you look like, and now maybe you don't want to see it. But, uh, yeah, same thing as kind of this, uh, technology is used now, also in the filming industry. Uh, so it, yeah, it saves time, I would say, and it offers new possibilities. Uh, so, uh, that this designer also can concentrate more on this creative and inspiring, uh, also interdisciplinary kind of working and not stick to the implementation and, uh, this routine work.

Jonathan:

It frees the designer up to, to do more of the creative stuff because they have a virtual assistant or a tool AI to help get the basic, the basis and the foundation and the framework that the designer can work more efficiently, quicker, and get really down to the nitty-gritty of the actual, like creative stuff. Right. If you want to age an actor, how old 75, 65 there's a 10 year difference. What, what does that have to look like…[inaudible]

Stephan:

Yeah, and because it doesn't take to take that much work. Why not try 65, 70, 75 and see what it looks like. It's a completely new opportunity.

Jonathan:

Yeah. You get quicker to the, to the vision without having to put in all the hard work. The hard work still needs to be done, but it's more of the designer is saying using the framework to make the, the design choice of which way they want to go in. Yeah. Got it. Yeah. All right. I'm getting it.

So, as an addition to the potential of AI, let's take a closer look on the field of marketing, right? So, um, a couple of years ago, Nutella did this really cool thing. And they used an algorithm and they designed 7 million Nutella jars all differently based on dozens of design patterns and thousands of colors. And they got an average post reach of 3 million, um, 10,000 videos were created by users and they could sell all 7 million jars within one month. I mean, this is, this is mind blowing, right? These numbers are crazy. Now, are there any other similar successful examples of how using AI and design psychology is extremely effective in marketing strategies?

Livia:

Yeah. Yeah. So I think that the Nutella example  it's, it's kind of the tip of the iceberg, because I think there are so much more potential than just making individual designs and like random designs and trying to sell them. But where AI is really strong is identifying every user's need preferences, so you can target them directly. So one amazing example is Netflix, and I'm not talking only about the movies that it recommends you, but I didn't know that before, but they honestly also target the design. So they're are not only choosing the movie. They want to show you, but they also think about, okay, I want to show you this movie, this person loves this actor. So I took, I take that picture of this actor that this person wants to watch the movie more. So that's a, that's a really fascinating way of how, okay.

Jonathan:

So I've been wondering this like private I'm like, yeah. I thought that, I thought that picture for the trailer when it first came out was different, but [inaudible] of the actor that I like. That’s AI in marketing. Okay. Stephan:

So, Netflix actually has really interesting approaches to try and figure out the way you interact with the website, with doing with their website and then try and transfer, like what they've learned from other users that are similar to you and also like make a profile of you. What's interesting for you. And then, yeah, as Livia said, if you'd like it a specific actor, they'll try and show you images of that specific actor in the trailer, or if you're into like romance and they'll show you like a romantic scene and make you like much more interested in actually watching the movie or series.

Jonathan:

Amazing. Amazing. So AI does help in marketing as well. I mean, could this be like a future trend with also like a AI powered AB testing? Or is this AI powered AB tests?

Livia:

Yes, it is happening. I mean, that's, that's how, yeah. That's how Netflix works. So yeah, I mean, in the past where you couldn't target or find out every user's preference, you just like looked at what the majority likes and then you addressed this majority somehow, but then you left out maybe, I don't know, 10, 30% of your potential users, but now you have the potential to target everyone without that much additional work. So you can find, uh, uh, every user's preference, needs and, and directly targeted and also use design to target them directly. Uh, yeah. So what they need, so that's, that's already happening. Yes.

Stephan:

It's actually really interesting, but it's during this AB testing phase, Netflix found out, like if they introduce a new alternative B and test that over a certain period of time, actually, and if it's then really better, all that time in between people haven't been able to use it and profit from this like new technology. So they decided to take it a step further and implement these algorithms online so that they actually learn while you, while you interact with the platform. So there's, um, like they do a trade-off between how much new things do they try to get feedback, how much you like it, or how much they exploit, like what they already know. You're like, what then they've developed really interesting algorithms to like tackle this trade off between exploration and exploitation.

Jonathan:

Amazing. Amazing.

Livia:

I mean this targeting marketing, it's not only happening on Netflix. I mean every social media you're using and the internet and Google, uh, I mean, you, you leave your digital footprint on the internet with everything you use, so they know almost everything about you. I mean, it depends on how much you tell them, but, um, they have a good profile view and that's what companies used to, to make this targeted, uh, advertisement. Yeah, exactly. So that's a huge potential for marketing and user experience.

Stephan:

And also really, really important for the individual to know what's actually possible today, what all can be done with your data, because if you just always click. Yeah. It's okay. Yeah. It's okay. Without thinking about it, you don't know what's in the future as possible. I mean, you hope that people won't abuse this information, but it's always possible. It's always something to keep in the back of your head.

Livia:

Yeah. And I think you also, with this targeted you, I think we should pay attention that we don't create this bubble so that every person only stays in his or her own needs or what he likes, because yeah, I think that has a lot healthy and everyone should also experience new things and get new impulses so that we don't stay in this bubble where we only, uh, yeah.

Jonathan:

I mean, that's, that's the ethics question, isn't it, you know, with anything, any type of technology that the whole ethics thing comes into play. And I think it, like, like you said, Livia with, um, you, you gotta be smart about it, you know, don't fall into, don't fall into the bubble. Um, of course, what did I do all last weekend? Of course I'd binge watched like a really cool series. Of course they did, but I also was like, okay, that was good. Um, and then I took a bike ride. I went outside, I enjoyed my life and I think, I think it's just trying to find that balance and on an individual term. Right? Um, and for some people that could be a downside to, to AI. So, um, one concern of a downside of AI that often arises is the fear that it may take human jobs away. Um, looking, especially in the field of designing, like we were saying before, um, is this a question designers should be concerned about? And if yes, in what way?

Stephan:

Well, for me, it's actually really important in what time span, like we said, what we're good at doing now is like assisting the designer to do their job better and terrifying. It really important that designers are ready and open to use this technology to be more productive, just to stay up-to-date with all that's possible using AI technologies. If we take it a step far further and think about like, what made the word look like in 50 years, then I say, okay, if we go more direction, this general AI prob um, generally I, it may be possible, but today it's not imaginable. So far, for the moment. I think we can say that the designers you'll most definitely you'll get a lot of help from AI to make your life easier. It'll help you be creative, but I wouldn't say that there's like now immediate, um, fear of losing your job. What do you think Livia?

Livia:

I think it's a bigger question of it somehow because, uh, it not only depends on the technology stat that will be possible in the future. I think technology wise, we will get there, for sure. I mean, we, we've done so many things. If you look at the internet, if you look at how we proceeded into last, I think 10, 20 years, I mean, I grew up with a, not even a smartphone, like, uh, I think the first mobile phones were developed and now we're having tablets and iPhones and everything. So I think we humans, we push further as hard as we can, and we want to explore new things and develop new technologies. So technology wise, we will get there. And, um, it's just the question of, how we want to address this, uh, as a society, uh, to maybe already now ask questions, where do we want to get?

Jonathan:

Yeah.

Livia:

Uh, so I mean, normally technology moves ahead and we humans just like, are following. So to say, and now we have to possibility to maybe shape this process already a bit like from the beginning or kind of the beginning part. Uh, and that's also what we want to achieve with our initiative, because we think we need to educate the society because it's going to impact all of us and all of us have to make the decision how we want to proceed. So, because now only like data scientists and people working in this industry know what it's all about. And all of the rest of the society is just living with the impact and doesn't even understand. And that's not a good basis to, to think about the future and what we actually want. So coming back to whether they will replace the jobs, it will depend on how society wants to evolve. So what we want in a future technology wise, yes. But if we like to design and to draw, if we want to do that for the rest of our lives, I think we will be able to work that way. You will do that. So it's, it's one part is the technology, but the other one is also the just society, the regulations we make, the decisions we make. So it depends on those two factors and I'm really excited and yeah, I'm really excited about how the world will look in 20 years, 40 years.

Jonathan:

Definitely. Definitely. So, um, I think, I think it's important also to think about the idea of human beings. We don't like to adapt. Like we don't like change, you know, so take that into mind that might slow down the process where designers go, this is a great tool and I'm just starting to use it. And I mean, I remember when I first started to play with Photoshop, I was like, Oh, this is so now I'm like, okay, I get it. You know, I remember when I first started to play with all these different design tools and it was like, okay, this is a tool. And then I found my own flow into it. And if there's an upgrade or a feature, I'm like, okay, great. This helps me. And for the things that, that didn't really affect me, I'm like, okay, that's great to have, but I don't, I don't really need it. And also, I think, like I said before, humans are they're adaptable, but they have to make the choice to change. And that's something that I don't want to say can slow down process, but it can, it can give us the opportunity to decide for ourselves if this is aligned with how we want things to move forward. So I don't think it's personally, I don't think it's going to take jobs away. I think we just need to see how it evolves and how we adapt or don't adapt.

Livia:

Yeah. I mean, I mean, and then we don't even know how the work like the work environment and the jobs will look like in general, 20, 30 years from now.

Jonathan:

I mean, 30, 30 years ago, when you said, uh, I'm a web designer or I'm p… I work at, I retouched photos. So like, I'm sorry, what, there was no job for that, right. Or I create explainer videos. Sorry, what? There was no, no such thing, you know, so true. Who knows where it's going to be in the future?

Livia:

But I still think it's important that if, yeah, if you want to kind of, uh, don't fall back, you need to get into this topic. You don't have to become a programmer or a data scientist, but just see what those tools can do and what they can't, whether the dangers, whether the challenges, but what are the advantages. So that's something I would recommend everyone just to dive into it and you don't need to study it. It's just, yeah. See, see what you can do with these tools and what the general ideas with them and just stay flexible. And yeah, this, this kind of adapting thing, this openness, um, yeah, towards such new technologies, but I think you really, you can profit from it. Uh, especially also in this stage, we are now you, you said the thing about retouching. I mean, there's this other feature in the newest, uh, Photoshop, uh, edition where you can retouch it automatically.

So you don't have to click on all the spots you want to remove. Uh, but still they, they also say it's not per, perfect. As, uh, if, uh, kind of a professional researcher does it, but it keeps, or it takes over to largest part. And afterwards you can do some very finishing touches yourself. So it, yeah. At the moment it doesn't take away the whole job or kind of the whole process, everything. It just makes it easier and takes the thing, uh, yeah. Or does to things that normally researchers or designers are kind of yeah, don't really like to do.

Jonathan:

Exactly, exactly.

Stephan:

I think what certain is, that it will change the job and how fast is absolutely correct, what you said, depends on how willing designers and people in general are to adopt these new technologies. But I think if we learn how to use them and leverage them to our, yeah. To do our best, then it'll drive us forward in general and make processes easier.

Jonathan:

Yeah. Beautiful. So I also want to like touch upon this before we move on to the next thing of how I said with Nutella. And we spoke with Netflix, with marketers who currently need the help of designers, like, uh, social media initiatives or post preparation. The marketers may be able to rely a certain way on AI to prepare basic illustrations and stuff and such. So is there the risk that designers may somewhat be replaced by marketers going directly to, to AI and not going through designers? Or is it still something that designers are needed for that final touch that, that we were just speaking about? [inaudible]

Stephan:

I actually see it exactly the way you said it, that like, it'll be much easier. There's a lower barrier entry, as Livia mentioned at the start. So if some marketer wants to just test out the certain idea, they can do it themselves, but I think to get the final edge, and also there's always emotion involved when we interact with content. So this is something the machine cannot quantify. We're not able to quantify this emotional aspect of it, so that are always be need for the designer to look at it and really figure out in detail what it means for us when we interact with it and really go into the details to take it to the, to the level level we're used to. And we have learned to love.

Livia:

Yeah. And I mean, in, in marketing, it's always the thing about, I think the story you want to tell or how you want to catch the user's attention. And I think that's not only interests, like, as we said on Netflix that you have this, uh, kind of what is he or she interested in or not, but normally you also have these kind of emotional reaction you want to create. So that's what, uh, Stephan, um, also said. And especially what Cleverclip does is like they’re trying to build a story around or yeah. They simplify a certain complex topic, as he said, and they wanna wrap this in a nice story. Maybe also trigger some emotions and this complex task can't be done by AI. So what it can be done is maybe if you, if you then draw the illustrations, some part of that, you can automize for sure. Um, but, but this whole, the, the bigger picture, um, that's something AI can not make or consider. Yeah.

Jonathan:

Yeah, exactly. You know, I think there's a statement, uh, the role of designers will evolve from being creators to curators. And I think, I think that like, like Stephan, how you’re saying the emotional part of it Livia the storytelling, you know, you can't, uh, AI can't do that or if they can, not yet. Um, and I think there is something to be said, like you said, the view about the, the, the Photoshop where you, the, the Photoshop, uh, AI feature does most of the blemish removal, but then the designer goes in and does the final, the final touch, you know, I think, and if this all goes to the idea of going, going from creators and transitioning into curators, and I, I kinda like that idea because people, people hire designers, not that they can design something pretty, but they design something that is in a specific taste in a specific way, in a specific genre that designer or that design agency, or, and of course you can go on onto Google and type in designers in my, in my local area and hundreds will appear, but only a few will, will really fit the style of what, what you're looking for.

You know? So I think it's, I mean, this is just my personal opinion and I'd love to hear your opinion on it as well. Um, how designers will evolve from being creators to curators. What do you guys think of that?

Livia:

Um, Stephan would you go first or I, um [inaudible]?

Stephan:

Go for it.

Livia:

Okay. No, I, I think that's, uh, that's a very interesting idea. And what you shouldn't neglect is the limited scope of creativity in general of AI systems, because we set that we are in this week, AI stage where an algorithm and an eye is trained to do something very specific. And researcher found out that creativity is when or when a human being is creative, different parts of the brain interact with each other that normally don't. So you really kind of, you, you touch upon very different knowledge or fields in your brain and you connect them. And because there may be two things that no one else would connect, but you do, it's kind of creative, you know, because you, you create something new in a sense. And because AI has this very limited scope, it doesn't have the data from so many different fields it could touch upon. So that's what I mean that AI is still very limited in its creativity, because it doesn't have this broad field of knowledge. Um, it can touch upon and connect very different things with each other. So I think that's something else that still, um, is a huge advantage of, of human designers. So to say, and yeah.

Stephan:

I don't think I can add a lot, you pretty much nailed it.

Jonathan:

It's beautiful, isn’t it?

Stephan:

Absolutely. So, yeah, like the most important point is that our, we train these models on data. So we are actually fundamentally limited by what data these models have seen. So it's really hard to go further from there. So if we, if we train a network to generate a human face and we, for example, only feed, uh, use images on the street, then all images that will be generated will be other human on the street and not in the forest, not behind the sea. So we're really fundamentally limited by this, um, like by, by the data we use to train our networks. Yeah.

Jonathan:

I love that the, the, the different parts of our brain, they make those connections and that's the creativity, right? You, you bring two different ideas together in your brain and you can't do that in, uh, in, in AI. It's, it's beautiful. It's beautiful. Yeah. Um, okay. Let's bring it back down to earth. So AI, I mean, it sounds great. I'm, I'm, I'm loving this conversation, but we also need to be aware that AI, like you said, we're, we're, we're still in the, in the baby steps of it and it’s in its infancy. Um, so should we rather think of it as augmented intelligence for now? And, um, and how will that, or how made that change in the future?

Livia:

Yeah, so yeah, I think that's, that's also something, uh, we hear a lot this augmented intelligence, because it's not this, this kind of general intelligence as humans, or as we define human intelligence is, um, but I think it will change. So, uh, I think we want to push this further and see what we can do with, uh, technology and research is going there. I mean, this general AI idea, so that it's not only a commented intelligence, but really this human level, artificial intelligence researchers are really intrigued by this idea. And very, yeah. They do a lot of research on that. So I think, yeah, we will get there someday.

Stephan:

I actually see it as well that we're like in, you can even call this weak AI age, the age of augmented intelligence. So if, as long as we're only able to like, solve these specific problems, then we're more like in this range of augmented intelligence and actually maybe to take it even a step further back, like in the fifties, when humans tried to say you're okay, when have we achieved artificial intelligence, there was distinct called a Turing test where Alan Turing proposed that if you have to set up with, um, three agents, so one person, which is an interrogator, one human, that answers questions at one machine that answers questions. Then you play this game of asking questions and getting answers for say a minute. And then you try at the human, like the interrogator has to say, if you interacted with a human or with the machine, and as soon as you can't distinguish that anymore, you've achieved AI. So earlier you really saw this as a very, um, very shallow thing, very shallow subject. And I think for augmented intelligence that's, um, yeah, no, that's just actually a too limited way of looking at it. And if we can take it a step further and go indirect, like general AI, then I think we can start about talking about this artificial intelligence for the time being, I would say, yeah, we're in this augmented intelligence realm. Jonathan:

We’re in augmented intelligence realm. Okay. I love that. That's super cool. So looking at this, at this digital transformation that we've been speaking about over the years, I mean, it's obvious to see that all businesses that ignored this trend, they're suffering, they're failing to stay competitive and especially after the pandemic that's going on. So can we expect a similar development in the area of AI? I mean, I, I spoke about humans are not always willing to adapt, or ready to change. They need to figure out their own way. So can we expect a similar development to the era of AI when it comes to this digital transformation

Livia:

So that that, uh, companies and people will be left behind, so to speak?

Jonathan:

Do you think people will be left behind if they don't jump in or the, or is it, or is it the question of, you know, we're already seeing it with Netflix and Nutella and, and Instagram filters, maybe they don't even realize it yet, That AI is part of it?

Livia:

I think it's more that way. I think it's, it's already in, in so many different fields. Um, and yeah, we don't even know it and that's the point. I think we should know it. And that's why we should educate, and talk about this topic that we are not only passively influenced by it, but also can take action and think for our own, um, how we want to interact with it. Uh, what's our opinion on that matter. So yeah, I think it's the latter. Yeah.

Stephan:

Yeah. I think so as well. And when we see like how our economy has evolved, like now the biggest companies in the world are all like tech companies that use AI left, right, and center. And they've really started to build up like these monopolies. Like they have really like huge infrastructure. They have endless amounts of data. And they're just like now on another level and can do much more. So I'd say, yeah, absolutely. Some, some companies, like there's only one real search engine company left in this world. There's, uh, maybe three or four big social media platforms. And to build something like that from scratch now seems like, um, it seems like an impossible job and yeah, I think that'll even [inaudible] transitioned into more fields. Um, there are like a lot of companies are really aware of this fact and are starting to use AI technologies for simpler tasks now and are really aware of how important this will be for them in the future. Um, yeah, I think we can absolutely see that, see that trend.

Jonathan:

Yeah, totally. I mean, to talk about designers and do designers have to adapt in a certain way? Um, and are they risking getting overtaken by designers who have already started to implement AI into their processes?

Livia:

I think it’s not too late.

[inaudible]

Stephan:

I actually wouldn't say either. I think it's really important to approach this with an open mindset and to be like really willing to interact with these technologies when we're confronted with them and to start to use them as soon…, as soon as possible. And yeah, I think it could be dangerous in the long run if you're just say, no, I'm not going to touch anything of this, but I would not say that it's too late. We're still like in this start, like this start transition phase where AI is finding its way into design and that we're even seeing now is the right moment. Um, the earlier you get into it the better.

Livia:

And I think it, it also, how AI evolves in the field of design, also depends on what society values, something that is designed, you know, if they, how open society is to accept that something a machine created is also something creative or something valuable, um, or whether they just like, okay, no, I think I want you to get something that a human being has made. So I don't know whether our society is already open enough to, I think Stefan, you said that, uh, some chairs and, um, bicycles were designed by AI. And I think society is still very skeptical, whether this is now a really good tool or if it's really that as valuable as if it had been created by humans.

Stephan:

Actually, this is a bit of a question of scale as well. Maybe like Nutella is a great example of where you saw people who were really into having something that's unique and was designed by a machine. But if you take it further, like an artwork that costs maybe millions of dollars, if you take it that far, then would you be willing to pay this for something that a machine generated and that you could actually just press “repeat” and which doesn't have like weeks of work behind it, or even months or years, I think that's, that's where this other kind of value that we human see comes in. And I don't think that AI will be able to replace replace that like this. Yeah. You know, the special feel of something that somebody put significant effort into doing.

Jonathan:

Yeah. I mean, you buy a Picasso because it's a Picasso. A Jackson Pollock, because it’s a Jackson Poll…, you buy it because of the artists because of the, the connections in the brain and how it comes together and the creativity. Um, I know that, that for me as a designer, that there's when I'm, whenever I'm like, “Oh yeah, I get to learn this new technology and [inaudible], there’s always that resistance. And then I actually do it. And then I realize it's like really good. And I get so excited about it. Right? So I know the resistance in me and I know what it comes up, but, um, I'm self aware in that, in that respect. But I know that when something new comes out onto the market, I'm always like, no, I don't need that, pushing it side. Right. So there might still be a lot of resistance when it comes to opening up towards AI for, for designers. So why do you think this is the case? Why do you think people are against it at first and don't want to adapt and, and what would be some advice that you would have to those, to those people out there about AI and leaning into it, opening up to it?

Livia:

I mean, you don't, I wouldn't recommend just kind of use every tool that comes out. So, no, I mean…

Stephan:

You won’t finish, I can promise.

[laughing]

Livia:

I didn't mean it that way. No, I just, what I meant to say is, it’s worth looking at the tool, like the tools that come out and really try to understand it and then decide whether you need it or you like it or not. So this openness and effort, uh, to learn something new, maybe just have a look at it, see how it works and then decide for yourself whether you want to use it or not. But I would recommend to be open and eager to learn new things because, um, yeah, new things will come and will shape our work and how we work. So that's, that's really important. I think. Yeah.

Stephan:

I actually think the most important step of this process is the initial, like the initial contact. You have to be willing to get into it and to make these first contacts, because that's what will be hard. But in the long run, you'll really notice how much easier it makes your life. So you have to be willing to make this upfront investment that the socket will most definitely make you slower because you don't know how to steer it, even what did, uh, or how to control it, or even what it means, what it's good at, what it's not good at. But yeah, I think it's absolutely worth just, um, yeah, testing out the boundaries and seeing what's possible and how we can make your life easier. Maybe something completely different for you, Jonathan than for me or for Livia. And yeah, it's just this, we have to be willing to go through this process of…

Livia:

And I mean, people are also afraid of these new things, as we discussed, they're afraid that, uh, those tools will take over their jobs or might be better than them. But I think if you learn to use it, it should only push and bring you further. So it won't steal something from you. If you learn to use it, to handle it. You, you can create even more and better things. So you, you should see it as a support and not as a, um, kind of a competitor. Yeah.

Stephan:

Yeah. That's really nice, we, we’re al.., we’re always better together.

Jonathan:

Exactly, exactly.

Livia:

yes, yeah.

Stephan:

Joint effort is always the best.

Jonathan:

You know, when you two were speaking about this, I just had a thought of, you know, when was the last time that I was resistant to change and to adapting. And every time I go onto YouTube and I do like product review and I watch people unbox things and how they work and, and, you know, nine times out of 10, they're always like, it's really great. It's a great investment. It's upgraded my business and I just go, wow, do I need that psychological, um, that like affirmation, like go for it. And I've wasted like two hours watching everybody say pretty much the same thing. Um, so that, that I, now I get it, you know, now I, I understand it on my, uh, in my perspective. Um…

Stephan:

So you think we should make more like AI product videos that you can see that it's really great?

Jonathan:

I mean, [inaudible], why not? I think, I think because, because we're, we're so used to going on to Instagram and trying a filter or doing all these things with the new, the new Photoshop feature, I don't think we re…, like I said before, we don't realize that it's AI, but it actually is, or augmented, um, intelligence. And, um, so why not do some decent product reviews of like, Hey, this is actually AI. Why not?

Um, we, we touched upon something that I wanted to bring back to before we, we, um, we finished the conversation and talking about, you know, ethics and emotional intelligence and, and principles. So, um, there's a need for certain principles when it comes to AI. So what would be some design, AI design driven principles that you think about, that you think might be good?

Stephan:

Actually, what, what drives these principles is problems we’ve figured out that our systems have, like, for example, biases, one of these big problems, like one crazy example of that has been going around social media recently this last few weeks is, um, translation. Like the Finnish language doesn't know the concept of “he” and “she”, so of pronouns. And then if you translate certain set sentences from finished to say English, then depending on the context, it will be translated as “he” or “she", and this is a fundamental problem because it's actually not what, um, yeah, it has actually nothing to do with the problem as such, but it's just driven by the data it was trained on. So, bias is one of these huge problems when actually our machines learn things which actually are not part of the problem they're supposed to tackle. Exactly. So what's, what's really important is that we're aware of that and that we also counter these, and that we also do research in direction of countering these bias problems and yeah, what concerning general principles. I think that AI should just, uh, should be human centric. So we should focus on how can it assist humans? How can we use it best for our society? And I think that's actually the most essential part. And we should always be aware of the risks that are involved with the new technology that we're developing.

Livia:

Yeah. And I totally agree with what you said and I, again, I'm coming back, it's I think those principles and guidelines and regulations, they will come when, when we figure out where we want to go with AI, because it's like, if you want to replace the jobs or kind of push the technology further and just go with it, then the regulations might be different, um, from when you say, okay, no, we just want an AI to support our work and not going any further. So it really needs this kind of discussion in society. And the first step is to talk about it and then to discuss it and make it approachable to everyone.

Jonathan:

Fabulous. All right. So let's wrap this up, looking at our main question from the beginning of this interview, how AI may impact design, what are the main takeaways now? What are some takeaways that you want to give our podcast listeners? How AI may impact design?

Stephan:

I would say there are three main, main points. The way AI will impact design in the short term. One is it can, it has like Livia explained earlier. It has the potential to lower barrier to entry. So if you want the new website, it'll take you a few clicks to just say what you actually want, what’s important to you, and then you'll get a recommendation. It'll not be perfect as we discussed, but it'll be a basis for further consideration. Second is that AI will be able to take over repetitive tasks and say clearly defined tasks, and then like take over part of work that is repetitive. And that's not like actually the creative process that we're actually interested in, when we do design. And a third aspect is that it can actually even contribute to creativity by giving new opportunities like new, new ideas, or maybe like just a different design that looks a little different may like, let us go down a whole other path and think of new designs. So I think these are the three ways in which it will impact design most in the short term.

Livia:

Hm. I totally agree. And maybe because I'm working as a photographer as well, I want to address the designers and give them something on their way. So, um, I think they shouldn't be afraid. Um, you, you should, actually designers, should be part of this discussion about AI and how it's used in their field and really get to know the tools and, and become part of this group or this idea, these technologies that will change and evolve and, and make use of it. And, um, just bring in their ideas as well, because we need to shape this future together. I mean, it's, yeah, it has an impact on all of us and we need all the inputs from every field everyone will be affected. So I, yeah, I just wished that, uh, yeah, just have also fun with the tools and, and really try to get in touch with them. Or, and, and I'm also very interested to talk to designers as well about this topic. So if someone wants to talk about it, I'm all open because we need this interaction with people working in that field. But also like, I mean, in the AI data science field, but we need the connection to where it is applied and how these people work to, to kind of combine our knowledge and make the best out of it.

Jonathan:

Most definitely because you could create AI and it could, it could be a great tool, but if nobody really uses it, then it's, it's pointless.

Livia:

Yes.

Stephan:

It was actually a really nice invitation to collaboration, Livia and what I would like to add here is that to be able to take part in this discussion a designer, must not be able to code or understand these mathematical models in detail. But I think it's really enough to have like a conceptual idea of how these algorithms work. Like maybe that we don't say do this, do this, do this, but more like have flexible models that are then trained on data and exactly. So that we just have to have like this high level of understanding and not like really in-depth understanding so that we don't have to be scared of, um, interacting with this college.

Jonathan:

Yeah. I mean, my takeaway from this and speaking with the two of you is two things, the, the idea of from creator to curator, I mean, isn't, isn't that, isn't that the goal like, isn't that, isn't that kind of like awesome because, you know, sitting behind your desk and doing the repetitive tasks that maybe could be done by a virtual assistant or AI, right? Where you can really just work on the concept and really curate something that's totally specific and specialized for your client is something that I go, I want to invest the time in that versus the boring mundane tasks. And the second thing is, is, like I said before using it as a tool, right? So having that as a tool, which frees me up to what you said Livia, those parts, your brain, where they connect, make those connections. That's where the creativity really lives. So being a curator hat, using understanding how my creativity is my own paintbrush. And like you said, Stephan of why would people buy an AI painting that can be replicated over and over. People want to get the Picasso, they want to get that, that individual thing. So it, it definitely goes from Oh, designer to, Hey, I have this vision and I know how to implement it quicker and have the tools and they know how to do it. Um, that's my takeaway from that.

Livia:

Very beautifully said.

Jonathan:

Oh, thank you. Couldn't have done it without the two of you.

So pointless science fiction, I think not. AI is totally a groundbreaking future technology and it offers so many enormous potentials in the design process. I mean, especially in the optimization and analysis of processes and using AI as a tool to support design, to work more efficiently, like I said before, that's what it's all about. So everyone, I’m very excited to see where AI will take us in the future. And after chatting with the two of you, um, I feel like we're in good hands. Thank you so much, Livia and Stephan for spending time with us today and for shedding some light into all things AI.

Livia:

Thank you also very much. It was a pleasure being here. Um, yeah, thank you so much.

Stephan:

Thanks a lot from my side as well.

Jonathan:

It’s great to have you two on the show.

Livia:

Thank you.

Jonathan:

And thank you to our podcast listeners for tuning into this episode of CCTalks - the Cleverclip podcast. Now to learn more about what we do at Cleverclip and how we create animated videos and interactive experiences that help explain a complex topic and inspire your audience with an idea, sort of like what we did today with AI, complex topic and making it more understandable, hop on over to cleverclipstudios.com. Thank you so much for listening and have a great day!