Tech & Design
Hi, everyone. Thank you for joining us on this panel today. And then our title is Tech and Design. We really want to enter into the knowledge and insight intersection around tech and design.
And the key role of design in business is growing rapidly, especially in digital products and services arena. And with the rise of generative AI, we are seeing big change and shift in design field. Just as like other industries and into today, we are super lucky to have an amazing panel of experts from the intersection space of tech and design. Tomo from Kontrapunkt, Japan. Kontrapunkt is one of the top design firms in Europe. And Jay from Silicon Valley, thank you for joining us from the US.
And he's an AI director in Figma. Figma is the most powerful online design and engineering tool maker. And my professor Miles Pennington from the University of Tokyo and the Todai is trying to build up the design new design department in 2027. So they will be sharing the insights and the real world examples, helping us to understand the future of design and AI. Okay, so let's start by asking
each panelist about the key technology trends that have emerged over the past year. Uh, Jay, thank you for joining us from the US. And I actually use Figma every day, 3 hours or 4 hours at minimum every day. And it actually Figma has completely changed how designers work, and I can. I cannot imagine going back to the file based, disconnected teamwork for. And thank you so much. And first of all, can you introduce yourself and Figma briefly. Yep. Yep. Yeah. First of all, thank you all for
inviting Figma and myself to the, um, to the conference. Um, so yeah, so talking about Figma. Um, to me, Figma is a technology based platform for product development, and the mission of our company is to help users and teams go all the way from an idea to a product. So we have launched multiple products to help that happen. From Figjam for brainstorming to
Figma design for building a design from your brainstormed idea to dev mode to take your final design converted to code, um and recently Figma slides to build presentations and then get feedback from from others. And with AI. Um, AI is becoming a transformational tool for all of product development. So this is something which we're very excited to to build and use in Figma. Uh, the way we think about AI. For us, AI empowers our users by giving them superpowers, both by raising the ceiling for professional users. Um, so an example is, um, when you look at designers, um, they don't want to do maybe like, uh, monotonous and boring tasks like renaming the layers. So we recently we launched an AI based feature which could take a design look at its whole content and then use AI to rename layers. Uh, we also want to make sure that
we give, um, assets and resources to the users at their fingertips. So we recently launched AI search, where you could either query with a with a text query or even draw draw an image, and then we'll find assets matching those queries from your list of assets. Um, we also see AI as a way to lower the floor and have more people use design. Um, so you may be a developer or a program manager who wants to build a design and share it with the leadership and then get feedback. So recently we launched two features like first draft, where you could give a prompt of a design, and then we'll build an early design so that you don't have to start from scratch.
Um, so yeah, so these are all things we're very excited about. And we're also actively working on a lot more AI based features, both for Figma design, but also for all of the other products. And as the field emerges, we want to work with all of our users, get feedback, and make AI features more delightful to our users. Thank you Jay. We've seen so many happenings or things around AI development in the world, right? And then this is my question. How have you changed your expectation or prediction on AI over the last year? Yeah.
So yeah, that's a that's a very interesting question. The pace of AI development has really accelerated in the last few years. So ten years back when I completed my PhD, that was when the whole neural network revolution and deep learning was just becoming popular.
But in the last few years, with the advent of data and compute with generative AI models like Llms, things have significantly improved. Some of the ways, like we have kind of changed the way we use AI is with these text models becoming really good. That becomes a new way to like interface and talk to products. So so one good example I like to talk about is like there is always a sequel language if you're a data scientist to to query databases, and that is pretty hard to write. I always have to look up references, whereas with AI you can just use a text prompt and then get the right answer back. So a lot of ways in which we
interface products would significantly change because AI has become significantly better for natural language. But recently what we see is it's also moving to other modalities. So generating very realistic images is possible. Uh, recently even like videos have become a lot more possible. Um, also with the foundation models, the other big change is you don't need huge data data sets anymore. To train these models, you could take a foundation model which was trained by one of these existing companies, you could have a small set of 100 to 200 examples to change it to the way you want. Um, so that again, would kind of
transform the way we start using AI more in all of these products. Thank you Jay. Okay. Let's move on to Tom Watson's part. Thank you for joining the panel today. And could you start by introducing Kontrapunkt and your world first, because I'm the representative director of the Kontrapunkt Japan. And Kontrapunkt itself is located in Copenhagen, Denmark. And we are the branding agency. And what we do is help the brand
to be much more global level. And yeah, also do like, uh, strategy for the design and execution of the design and implementation of the design. That's what we do every day. And of course, we do use Figma every second in our company too. Thank you. Yeah. Kontra-punkte is known for its excellent design strategies and graphic designers team. And, uh, have all the designers and strategists in Kontrapunkt started using AI in their everyday activities or work. Uh, really? Good question.
Well, in the kontrapunkt, we have this, uh, three value on college craftsmanship and curious, and all the contracting people are very, very curious. So in the past two years, we have been using all the I like the photo generative AI and image generative generative AI. And we actually made a summary that it's not there yet. Sorry. And we actually has that I you know,
there's a risk that I can do the designer's work, but we conclude that it's not still there yet. Of course it's going to grow. But so after we tried every tools, we decided to still not to use it as an design direction, but we use as a design guide. As they said, we can use that as a first draft of the design.
And we as a human being, we are trying to overcome that design. And but the one thing we use are relatively or openly is the photo generative AI, because it's very, very easy. If you put the some of the images, you can create another photo image, and it's much easier to go and go outside and take the photo with the cameraman. But that's also as a draft.
So for the final delivery, we are going to actually do the photo take because it has an actual photo take, has a much more sizzle, and we trust the photographer that you can make the better than the AI. So yeah, that's our stage right now as a design to use AI. Interesting. So the core components or core activities of designers are not replaced by the AI. But not yet. Yeah. Not yet, but, uh, same question. Um, have you changed your expectation on AI over the past year? Uh, good question. I think it's a yes. I think our company has been changing
our perspective on AI, and we are more open to AI, and we actually embrace the development of AI. I know some of the Japanese company doesn't really allow to use the AI or not, but I think we should grow together. And with the AI we can also designers should grow together. And one thing I just found out
while I was looking at the designer how they operate the AI to make the accurate prompt is the most skilled, actually. So maybe designer's skill will be different in the future like they have to prompt correctly. Yeah, I feel exactly the same thing. It is more like the metaphor is quite similar as like a relationship in between art director and design team, because art director use words and communication to direct the design quality, right? So the words or communication kind of narrative communication is quite important for even for designers. Yeah. That's interesting. Thank you. Nice. Okay. You're from the education sector. And then please introduce, uh, Tokyo University activities on design and yourself. Thank you. Kenya. So University of Tokyo, my position there is a kind of creative catalyst. We're exploring different ways
of bringing design into the culture of the university. The main start point that we've had in the last seven years has been a design lab that's positioned in the heart of a science and technology research institute. So the work we do there is trying to bring design methods and processes directly to researchers involved in very frontier engineering or technology development. Maybe an example in terms of AI, of the type of research work we do is we have a project called democratizing AI. So it's the university of course, is full of like fascinating AI development, not just generative AI, but all other, uh, aspects and branches of it. One of the things that maybe
design is good at always is bringing the human element. So connecting the university with the general public to discuss what is AI. Introduce it so people understand its benefits Fits and and and problems to and then bring them into the conversation like how can we use it. So it's only a small project, but I think it's a nice example of how design can get involved in the in the kind of research development part of, uh, of AI. Um, so that's. Yeah. And the second thing I do is
we're evolving a new design program at the university. So University of Tokyo is not doesn't have a design program yet, but it's for many reasons, seen as an important future for the institute. Have you noticed that the AI impacted on actual student or researchers activities? Yes, of course. I mean, there's a huge fear of wave of fear has swept through education. It's like, who's doing the work these days? Like, is it the student or is it this some clever AI and I the basic, um, position for me is that I'm very excited about it as an educational tool and as a tool for students to use, of course, both creatively and in all aspects of of education.
But where educators now have to get up to speed is they or we need to lead in terms of how students engage with these, with these tools. At the moment, I think the whole sector, the education sector is going through quite a kind of tumultuous period of trying to figure out how to engage, how to control it, how to leverage it. So yeah, it's it's huge. But I think and it's still early stages. Right. Like Tommy says, it's not quite there yet. And I think it's that's true of
course, in education. But one of the really exciting things in design education is some of the tools make it really inclusive for people to get involved. So people who can't express ideas, either visually or in other ways, can suddenly have a way of bringing things to life that's really interesting. It's still pretty basic and cliched, but it is really inclusive, and that opens up the world of design and creativity in In New Ways, which is, for an educator, very exciting. Okay. All right.
Same question for you. Have you changed your expectation or prediction on AI over the past year? Yeah. Well, it's kind of gone from zero knowledge to massive excitement, really. And I think for, you know,
we're a we're a creative lab inside a university for us, just actually brainstorming the future of how AI can be utilized. The other day we were, you know, we're planning a new course, and part of the new course is that a student should be able to access any knowledge inside the university or globally, basically. And of course, now we're trying to practically deliver that. So how many lectures can we get to deliver it to? How many students then? You think by 2027 when the target to open it is, maybe a student can just call up, you know, a custom made lecture on a subject they want delivered by a person that they want by this AI platform in the most interesting and engaging way for them. You know, it could absolutely revolutionize the way students engage with university and allow educators to do what they really best at. So I think it's yeah,
it's a revolution that is hopefully coming soon. Yeah, yeah. You know, among, uh, all of you, all of you guys, I can get the sense of last year, we we have, you know, wow news. It's AI is new things, but we really get in touch with a more reality side of the AI. And I use the AI as a design or creative tools. Yeah. Thank you. All right. So let's move on to our second topic, the AI's potential. I'm really interested in how AI is transforming creativity itself and designers and design industries. and I'm asking each panelists to
share your visions for the future of design in relation to AI. So, Jay, you directing the AI development in Figma, right? And then how do you see AI will change the design process over the next five years from concepting to prototyping and final production? Please share your vision. You know, these kind of design process, uh, you know, the image of, uh, for the five years future? Yeah. Yeah. Um, yeah. That's a that's a really great question. Um, so as I was telling before,
um, AI is a very powerful and transformative technology for all of creativity and for product development. Uh, to me, there are different waves of changes that would happen both in the design process and the product development. Some of the early changes we expect are.
Um, like automating the more boring tasks, uh, making sure that designers become kind of, like, more productive. They're able to focus on things which they are excited about. Um, for example, like, as I was telling before we launched some of these features recently, um, in Figma. I like the rename layers, which is something most people I talk to ah, don't like to name their layers. And now I could completely do that,
so you don't need to worry about it. I would take care of it. Um, that's like the first wave of changes we are expecting to happen now. The next would be to again, having data at the fingertips when you want to engage in the creative process and when we need inputs for that, we want to make sure that it's all available at your fingertips. This could be assets we provide through search. Maybe bigger fragments, maybe components in a design system, which a designer could use to quickly build their design. The other aspect in which we would
see AI changing the whole design process is bringing more people in. Um, so before, um, to build a good, high quality design, you need to be a designer. And maybe if you are early on, it's very hard for you to get in. So we want to lower the floor, make sure that more people could build early designs, get feedback, participate in the whole design process. Um, those are, I would say the, the, the initial wave of changes, but there are also bigger wave of changes we expect. Um, so, for example, if for people who are familiar with design systems, uh, when Figma introduced Design System, the biggest change that happened was instead of a designer thinking about pixels and colors and rectangles, they could now think about bigger concepts like, uh, maybe style and typography and so on. So for people not familiar, design
system is a collection of components you could use to build design, and AI is going to make a lot more change in the same direction. You can now think about larger and higher level concepts, and you can think of the whole design process in terms of that. Um, one of the extremes is like, yeah, the whole like using first draft where you can give a text prompt, get an early design, so you don't need to build all of that things. I would take care of it.
You could still add your craft on top and make it change it the way you want. Um, that is something that's exciting. Um, the other longer term change I would expect is today, if you look at Figma products, we have one product for each area we have like Figjam for brainstorming, Figma design for designs, dev mode for development. But I could also help you like transcend across these boundaries and then go all the way from one point to the other.
Maybe you could brainstorm some of your ideas in fig jam, and I could take all of that and convert it to a design, or to a slide you could use in a presentation. We already do similar things today where we can cluster ideas into in a fig jam into like different groups, and then you could look at each of them separately and the whole thing would happen in other products too. Maybe you build a design. I would build the code for it, and then you get a first draft of the code, and then you can start working on it. So that would be the bigger change. We would see where I would transcend across products. And you don't need to think about different people for different products.
But I could start at a point and take you to the to the end point. Um, so yeah, lots of lots of changes we expect, um, with AI in the whole field of creativity and product development. And yeah, I'm very excited to build those features, work with all the users, get feedback, learn along the way and make it better. Thank you. Do you have any comment on that? I was just going to actually ask. So I suppose I sit in the 1 to 4 position in education of not having to make and sell anything, but just train people. And I was wondering how Figma or rather these sorts of learning platforms to, you know, explore activities that we can't yet imagine, you know, like, I think that's what's really exciting about how these things are evolving.
They're filling gaps at the moment. But what's beyond that? Like, how do we how do we find out those, those activities or tasks or things that we can't yet imagine being done by something like AI, that actually the really revolutionizes the way we work. And I was wondering how, yeah, commercial entities like Figma and design consultancies are thinking about that.
Yeah, I think that's. Yeah. Yeah, that's I think what uh, Jay said like the lower the flow of the user for the AI, that's kind of scares the designers out. Because what if our clients did all the plumped and bloat? This is our logo design that AI says it's good. So can you overcome this design?
Of course. That could be happening in five years. And of course, to, uh, to overcome, you need a human brain and hand and all the ideas come through from like the one or 2 or 3 designers. So I think our designers or the branding skill should be much how would you say crafted than the human than the AI? Because I think as far as I see, every human has done little. Every designer as a human have a
little bit of the edge of them, of themselves. So I think maybe I can also demonstrate that one. But maybe the one human being can also do that and overcome those challenges. So yeah, that's kind of scares me
that all the client bringing the AI. Yeah, it's kind of true. True voice from design firm. Jay. Do you have some answers to them? Yeah. Yeah yeah yeah, totally. So yeah. So so maybe, maybe talking to the first question about like, what are the bigger pieces which I could help. Um, I think we're still like, learning as a community. Um, so we've been taking small steps,
getting feedback from the users and then figuring out how I should be evolved. Some of our key tenants are making sure that, uh, things which could be automated. Can be automated so that especially in a creative process like design, designers can focus on the creativity part, on things they are excited about, don't have to worry about doing things which are not interesting to them. But yeah, I agree, it's still like a lot of learning to do. And we're taking small steps and then
understanding our users and then um, going in that direction. Uh, maybe coming back to the second part, like about this whole maybe the risk part of, um, the maybe the, the need for the craft. Um, the way we see AI evolving is there is still a lot of craft which is needed in design, which the designers bring. So we mostly think about AI as a way, as a tool to complement designers, not to replace them. And the craft is still going to be a very important component for design, which the designers would actually carry. So yeah. So we're kind of more excited about, uh, getting the the boring tasks automated, having designers focus on the craft and creativity. And even in the coding side,
it's exactly the same. Uh, you could maybe use AI to create a lot of code, but then the important pieces are like, what is the right software system to build? What are the constraints which the human has, and how do you make sure that those could be handled? So some of the questions which need more craft, those can still only be done with with like designers and developers with all of their expertise. So we think of AI as a way to like, complement and take care of the more boring and tasks which like, want to be automated. I just want to throw in a point which is about the fear of, uh, creative work being taken over by AI, and it reminds me of when, uh, digital photography came on. Right?
And everyone thought that would kill photography, but it hasn't because I think ultimately the proportion of great work is still the same. So the volume of kind of work in the world expands with these tools, but the proportion of excellent work stays the same. So that's where the professional and the craftspeople come in, that they will always be at that excellent stage. And the world has will accept other qualities coming in. But I'm yeah, I think certain other jobs may be taken over. But in terms of creativity, I think there's plenty of stories from the past that we've seen where these tools have added and not taken away. Yeah, yeah. I've got the related question to
Tomo. The, the so many designers are worried about the AI replacement of the human work, but the, um, as we as we discussed, the AI can take over repetitive work or non creative work. And so it allows us to focus on more meaningful work. So have you noticed any any tangible
example happening? Kontrapunkt. For example, we can actually use, uh, AI to research, um, summary creation, for example, and then we can actually shrink turnovers into kind of 10s for that, for example. So do you have any? Of course, I think it's much more uh, strategy part, we are using AI, especially me, as using AI a lot to, to uh, how do you say, uh, shorten our strategy part, like especially the research part. So if you prompt all the, like the strategy work or any, uh, questionaries or like, say, quotation from the client, etc., we can actually make the quick summary with the AI, and they have much more knowledge than my brain.
So I have to trust the AI knowledge for making the strategy and summary. But of course I will add of course my taste on it. So that's what I do. And I think with the AI, as I said in the beginning, I think designers are taking a lot of risks. The task by AI that they can do the initial research of the similar logo. If they create one logo,
they can research if there's any like similar logo with the AI. So I think a lot of the task has been taken by AI already. But, uh, as you guys also said, the craft part is still remains. And I have one really big good example of the design that still remains. Uh, our founder created the Lego logo
like 40, 50 years ago by his hands. And it's still there. And we have been trying to change it so many times, but none of the design is beating that design. So it's still there and it's handwritten. And yeah, so I think as you said, we still have a cloud by human touch. Yeah, yeah. That's interesting. And in my studio, actually, our designers and engineers use using AI every day. And I noticed that the speed of
iteration was doubled and tripled. So that that means that the, for example, within the design thinking process, iteration speed is a kind of core crucial parameter to boost up our innovation innovative thinking. So that the I think introduction of AI is going to change the process of design thinking from the bottom, because that the the limitation of time and resources are the actually the parameter to define the design thinking process.
So I think within five years, I think design think new type of design thinking will be established in a way with I miss you. Tokyo will start the new College of Design in 2027. This this is this is amazing. And and then you are the the person
to lead all the program planning. And you are aiming to educate new type of leaders, new generation. Right. And uh, as you said, naturally, I will be the fundamental part of education, you know, in five years future.
So do you have any vision to use AI, uh, in the future? Education? Uh, yeah. I think that question has kind of two directions. One is how we use AI as an educating tool, which I think is very exciting to think about. You know, there are so many elements of teaching that I think could be done more interestingly and more engagingly by AI in the future to allow people to inspire and lead students in in the most powerful way that they can. So I think the role of educating
is going to go through its own revolution. And it's education is, yes, is a slow changer. So I think, yeah, we as universities and colleges need to speed that process up. But the second is like,
how about using AI as a kind of subject within the course itself. So we say we're opening a college of design, but in truth it's not it's not really a college of design. We're trying to change the type of graduates that we create and to make people more aware and engaged, capable of understanding issues and opportunities and giving them the tools to create ideas and to to deploy those ideas. And design is kind of the discipline that's great at doing that. So our mission is to teach everything, everyone about anything, and then give them the opportunity to create their own solutions and actually test them, employ them, and create work that we can't yet imagine and hasn't yet been seen. So design is the powerful discipline that enables that. And of course, AI is one of the
amazing societal opportunities, I would say, and has its issues. That is going to be a subject of close debate and focus of projects. What students will do, I don't know, because that's what I love about design, is that they should be doing things that we cannot yet imagine. And I'm yeah, I'm looking forward to giving them the platform to enable that exploration. Yeah. I'm very curious in what kind of what type of future leaders are, you know, emerging in the, new context of education with AI. And then personally,
are you focusing on individual genius type leaders or more collaborative collaborative type leaders? I think it has to be collaborative because what we there are many great design schools out there looking at product or service or, you know, communication design. We're not trying to be another design school. We're trying to create people, you know, who are more change makers, who can understand the problems and issues and opportunities in society and propose realistic ideas. People who will who will shape the world around us. And there are many interesting
things out there. There all tend to be wicked problems. They cannot be resolved by one person alone. So for me, the future is about collaboration. Through collaboration, you not only can create more diverse ideas and solutions, but hopefully you can deploy those ideas.
I've been in education over 20 years, and I think the best projects I've seen walk out of education and become startups and successful businesses are now always team projects. They always start that way. It's a much stronger place to be. The age of the of the superstar designer is gone. Okay. Thank you. Jay. Um, we've got interesting thoughts from, uh, Miles.
And then, uh, do you have any words around the relationship in between new type of creative leadership and, uh, I. Yeah, yeah. So I think, again, um, the way I think about creativity, AI is going to like, accelerate the creativity a lot more, uh, because now the creators, like, designers, do not have to do all the things, but they could leverage, uh, AI to automate some of the less crafty things and then focus more on things that need more craft. Maybe understanding the problem or bringing in the human touch. So I would say the right way to build the next set of leaders would be to, of course, have all of these AI tools also available in the fingertips so that they could decide the right way to use them. So there's always this worry of like, hey, AI is bringing a lot of change. Should we say no to the change or should we embrace it fully? To me, a conservative embracing of the technology is the right way to move forward.
So we definitely want to make sure that new designers, new developers have all of these tools at their finger fingertips so that they could focus on the more creative things, the harder problems to solve, and use AI to complement their work. Thank you. All right. We are getting to the last part. and then the ethics or the risk of AI part. And there are a lot of different opinions around, uh, potential threat of AI or risk of AI. So I'd love to hear each panelist perspective on this topic.
Uh, so start with Thomason. Kontrapunkt is known for the very strong contribution to sustainability issues or ethic practices, uh, in, in European nations. So what kind of conversations are happening in Kontrapunkt around AI or new technologies? Thank you for asking that. I think we, as I said in the beginning, we are much more embracing the technology and I think we are much more embracing the AI. And we see that, uh, actually not the risk, but the challenges are there because I think every, uh, AI I solution will maybe look the same in the future. So as an design agency, we can help to create a much more, uh, friendly or matching, uh, interface for them or UX for them.
So that's what we can challenge to the older AI or companies or technologies. And I think one, uh, our clients are much more becoming demanding by using the AI technologies. But I think we, since we are embracing all the technologies we are now, have a position as a creative technologist. We don't call it as an engineer, we call it as an technology, uh, creative technologist that they can actually, uh, modulate or make the 3D mapping of everything that, uh, clients are asking for.
And one of the good example of the whole state is the energy, uh, Are solution companies that are making the renewable energy, that they are asking us to create a map of the what they do and with the location. So I think in the past we can't really make that happen. But with this, creative technologists in like one second, he could make the 3D map with the AI and put the all the important initiative of that company in like few days. So I think our our potential and capability as a design agency with the AI and technology is like much more becoming big and enabled and much more accessible to the consumer to, uh, in terms of the design ethics or kind of, uh, consciousness on the sustainability or human rights or do you have any? Uh, yes. I think it's, uh, with the design, I think we can make it much more accessible to the consumers. As an designers, I think we have much more responsibility to make design accessible to all the, uh, human being with the AI and technologies.
So I think we are also keen on that area. All the designers, especially Denmark, is known as the most sustainable country in the world. Yes. Thank you. Um, I heard some, um, planning around new design department to Tokyo University. And then one, uh, probably quite important component of education is ethics. And the, um, this rapid growing technology and situation, I think new leaders should have a kind of a strong, uh, spying to understand what's happening and manage what's happening by.
By themselves. Right. So could you tell us about the how are you going to educate new leaders, new generations in terms of ethics or risk management of technologies? Yeah, it's it's a really good point and kind of core to the new College of Design's philosophical approach is not only is it about, you know, education and research is about authentic and, you know, honest work and creating knowledge that is of of your own generation. And of course, AI in particular has challenges in terms of, you know, understanding authorship and, um, how it's really created. But what we have set out or will set out to do is to try to give students a much firmer foundation in terms of ethics, justice, equity, understanding the world around them, and the differences and the privileges that people have and privileges that people don't have. We want to put that back at the
heart of education. So people have a much kind of firmer understanding of how things work in the world, not just AI, but also an ambition to try to resolve some of the problems that those differences create. So I think it's always been part of, you know, a university's role to inform and encourage students in that in terms of working honestly. But there are so many more complex things happening now. We need to put it right back at the
heart of the way students think, and I think it's part of their mission is to resolve those challenges in the future. So maybe a small one is, yeah, the problem of kind of creative authorship in AI, but I am sure will lead to many other inequalities in the future, and it needs to be the highest priority to resolve those differences. Thank you miss. Lastly, Jay, I'm sure Figma has ongoing discussions internally about the potential risk of AI, right? So what kind of conversations are happening at Figma and what's your thoughts on this these issues? Yep yep. Yeah. Thank you. That's a very important question. Um, so yeah, so as I was telling before, AI is a very powerful and transformational technology, but with any other powerful tool, we also want to make sure that we build it and use it very responsibly. Um, so we spend a lot of time
inside Figma thinking about it the right way to build it in a very transparent way, and the right way to grow the AI features. Uh, we go at go to great lengths to make sure that the privacy of the users are preserved. We also use best in class security features so that all the AI features we build are secure.
There is also this other aspect of quality because when you build AI models, we also should make sure that it's fair, it's representative, and it's also diverse. And it should also be accurate, like it shouldn't have hallucinations where it gives the wrong answers. And then it kind of affects the the outcome. Um, so we believe you should, uh, build the models slowly, evaluated rigorously, uh, launch it slowly, get user feedback and then improve. Um, there is also this bigger question of how I should interact with the users. And our belief is AI should give
superpowers to the users and complement the work which the users are doing, not really replace it. And the first feature which Figma built is like a good example. It helps you, with the blank screen problem by giving you an early design, but there is enough control for you as a designer to add your craft and then make it better. Um, as AI accelerates, um, we believe that craft would become even more important. Um, so designers would have who are really good would be like even more important, and they would be kind of differentiated compared to everyone else. Uh, I would help increase the pace and the quantity of work, but the quality would still be with the humans. Um, so yeah.
So I would hopefully help the designers move quickly and then build more creative designs with, uh, with more craft, which are emotional, which kind of take into account the real, like, kind of building the right emotions, which we want to build. Um, so, yeah, so it's an exciting area, but we should also make sure that as a community, we learn more and also set guardrails along the way. Thank you. Jay. No. Okay. All right. We've got, um, 15 minutes for the floor. So, uh, yeah, let's open the floor to for questions and comments and uh, uh, audience, if you have any questions or comments, please raise your hand and introduce yourself and keep your question concise. Um. Have any one, two. Okay. So let's set these two, please. Thank you. Steve Monahan from Finn. Meri. Uh, just a question.
I refer to, uh, I as usually amplified intelligence because it always begins with the human, even if it's a prompt. But nonetheless, in Japan, a lot of the design improvement we do is process driven rather than customer experience driven. And yet we talk about how important emotion is. And and that's a core part of design. How do you elevate and move away.
So if you've if you've look at coming into Japan visit Japan app is process definitely not experience. How do we change that in Japan. Interesting. All right. We can get a second question. And after that we are going to answer okay please. Hi my name is Tyler. I'm a current MBA student here at Globis. Thank you everybody for the great discussion we had today.
I know a lot of I read somewhere last year that the amount of energy it takes to produce one generative AI image is about equivalent to charging your smartphone. So I'm curious as to to what extent each of you organizations are considering the environmental impact of implementing AI into all of your endeavors. Thank you. So for the first question, do you have, um, any words? I guess one more thing that I had to bring like more experience design into into design tears or where the process of that in Japan. Maybe that's a professional designers viewpoint. Yeah. You mean experience design? It's, uh. Good question.
I think we are the like the the foreign design agency in Japan are targeting the Japanese client. So I think it's, uh, it's still a learning process for us to do some experimental design, I think, uh, yeah. We still don't have an answer for that. Can I can I put it more bluntly? Why is it does digital experience design so poor in Japan? Is that is that is that compared to Europe or other places? Yes. I think uh, the first uh, I can't really tell the bad things about the client, so I'm just trying to find a word, but I think it's about the space. I think for our design, like as a
foreigner or as an Denmark designers, we actually appreciate the white space. As a designer or. But what Japanese are people? If there's a white space, we want to put them much more information. That's one. Yes. Counter that. I used to run a consumer bank here, and I heard exactly that thing was told to me. You don't understand the local market. We love all this noisy space.
We went an a, B, tested it. 100% of customers preferred the simple to the complex. Right. So I don't buy the story just to be as blunt as as we do in the southern hemisphere. Yeah, I have a counter answer for that. Yes, of course the consumer prefers the much more simple and white space design. I need to ask to cut the video
from here. But like what? The Japanese client. They are much older. We as a Japanese, we are much more inclusive as an organization. So we have to ask for all the functions input to create a one good design. So of course all the all the functions business function will say please put this one into the front page and stuff. So I think there is an uh decision making processes are losing in some, you know, points. So that's why we have to put a lot
of information in the front page and the second page and third page. That's why the experience is missing in our digital design. That's what I can say. Sorry. May I just, uh, probably we can go for quick. So I went through all those battles in Japan and we grew the business 25% in six months straight after we went. Simple. How can I help that problem then?
Yeah, like I management tool maybe or something. I don't know. It's like yeah, maybe these are. Well, these are like the craft of design that I can't help, you know, like negotiating and changing that approach. Isn't it so? Yeah. That's a very good point. Yeah. Okay. Um, the second question. Can you can you repeat the question? So a lot of generative AI use in design requires a lot of computing power. And I know there's been documented
cases where a generating a single image, as I said earlier, requires as much energy as charging your smartphone. Obviously, this iterative process is going to require a lot of kind of I didn't get it right the first time. I need to put another prompt excessive iteration, and I'm wondering to what extent each of your organizations are considering the environmental impact of AI. As you implement it more and more into your everyday work processes? Yeah, that's a very good question. I think our kontrapunkt is actually b-corp satisfied the company. So I think we are trying to use this power or energy every day in our office.
So I think one of the decision that we decided to not to I of the image generated every day is that part we are trying to consume the less energy our company. So that's why. Most J. J. Do you have. Yep yep. Yeah. That's a that's a really great question. Um, so so I would say the, the if you look at either the training of these large models um, or even inference to like, get the images as we were discussing before. Um, of course, it takes a lot more power compared to some of the older machine learning techniques, but that is also because the field is fairly new and there is a lot of work going on now to make these models smaller with distillation, even when you're training like instead of training from scratch, you start with the latest model and then train on top. Um, so there is a lot of work going on in the whole field to make these models smaller. Also like have more hardware for
inference, which is power efficient. Um, so yeah, so, so a lot of work would happen in all of those areas. We at Figma are also looking at reusing these models as much as we can, so we don't have to train from scratch. Um, the one big area which I'm very excited is the whole growth of open source models. Um, so for example, meta Media recently released llama 3.2, so you don't need to like, do the training from scratch.
You can start from the best model available today because it's all becoming open source and then start from there. So yeah, a combination of reusing existing models during training. Bunch of techniques to reduce the compute during inference, like doing things like distillation, quantization would all help us reduce the consumption. But again it's a process and the field is moving quickly. So I would actually expect more
progress to happen in that area in the next like six months to a year. Maybe. Just a really quick point from me just on, in terms of a kind of new wave of thinking that's going through the design industry and particularly in education, we've moving away from, well, there was a long campaign to get people to think of design as being human centered. You know, humans are the center.
And now there's a move to think that design is should be thinking beyond human or more than human. So human is just part of the complex ecosystem that a design has impact on. So I think the questions like that around, you know, the kind of hidden aspects of AI are those that need to be brought to the surface by that attitude. It's no longer good enough, just that my users satisfied, but that the whole ecosystem is somehow thought about. And the human is just one touch
point in that. I know that's like maybe overcomplicating many designs, but it's a way of thinking, a mindset that's important to for the next generation of designers to have at the very foundation of their way of working. Thank you. All right. Uh, this is Shane GU from Google DeepMind. So, um, in software engineering, the kind of space I think, like, there are a lot of, like, entry level positions in kind of programming are currently being taken by like I. I'm kind of curious what's sort of happening in like design kind of industry. In fact, a family friend of mine, she was like in like Disney, like sketching. Apparently the like, entire
department got laid off or something. Another kind of more kind of high level questions like, are the designs always subjective or dynamic with respect to humans, or are there some kind of objective metrics on design that we are all converging into? Like is there something like golden ratio, like more general version of that conversion into thank you, maybe Mouse Tomasson. Maybe I'll pick up on that second question, because my answer would be I hope not, because actually, I think that, uh, one of the, one of the future challenges for designers is to kind of respect a wider view of the world that they serve. And I think with, uh, you know, global manufacturing. We've become a bit kind of constrained. And, you know, in terms of style and the way we use things, everything is a bit kind of ubiquitous and unified, and it's much more interesting if we take every cultures and peoples viewpoints in mind and we and make more creative and interesting range of products and services and things that way we live and policies.
So I hope not that hope we're not converging. That would be very boring world, and I think it would be much more interesting. And there'd be a need for many more designers if we have more diversity and inclusion in the way that we think and produce. I think I completely agree with you.
And I think as a designer, we used to have to have our personality and age. So I think it's, uh, yeah, it's not going to be that not going to be happen, I think. J yeah, I can add some thoughts to so yeah. So about the first question of maybe entry level software jobs are no longer being there.
And if it's true in the design space, um, I kind of think of it as a much more like optimistic, uh, change that is happening. Um, now, given that I can do some of the boring entry level work, uh, developers can now focus on maybe the bigger system, like what are the actual constraints which the system has to take care instead of thinking about what should be the code you should write or what is like the how, how to fill the class, or maybe how to write the test. And um, that is the same pattern I'm seeing in the design space too. Um, there are lots of like, boring things, which is taking designers time, uh, which designers are not excited to do. And now I could automate that. At the same time, designers can focus on the creativity and craft. Um, and also for many of these companies where it's, like, extremely hard to hire designers, you can actually now get a lot more things done with the current set of designers by using AI to automate all of the boring things.
So I would expect existing teams and companies to become a lot more productive with the same amount of workforce at the same time. Also giving uh, maybe the the entry level software developer like more time to focus on more interesting things and grow their career. Since I would now do some of the entry level boring things for them. Um, so yeah, so I see the change as a lot more optimistic happening. And maybe a quick thought on the second question to on the creativity in design. Um, yeah. I still feel there is a lot of craft which comes into a design which which differentiates your design from maybe something that's already available. And it's very hard to have an
objective measure for that. So there is a lot of subjectivity and human craft that goes into really great design, and I don't think I would automate that in any of the near future, but that's a great question. Thank you Jay. Okay. We've got only 15 seconds to to finish this up. Okay. Um, I actually did a exactly same topic last year in this G1 global. And then I felt the quite a, you know, jump from the one year ago to current situation. Um, the AI is a kind of tangible
things and, uh, interesting to and then gradually changing design and design industries bit by bit. And, uh, thank you. Thank you, um, so much for joining us today. And a big thank, uh, thank you to our amazing panelists.
And let's give them a big hand. Thank you.
2024-12-05 00:11
