#129 How to become an AI designer
27:30 Share

Even if you don't work directly with AI interactions, maybe you use AI in your work and that would make your skills transferable to roles that need to solve AI problems. My prediction is that this role will eventually disappear because

all designers work with AI in some capacity. So we will all have to think about how AI is just a new tool in our toolbox. It's a new technology that we can use to solve the problems. The key here is that you're still in the driver's seat. You're at the center of this process. You design it and you decide what are the parts where you need to do the thinking. You are the one who decides what comes next. What is the outcome that you expect?

expect at every stage. So you're still in control. Even if AI is replacing some parts of the process or supporting you in some parts of the process, you still have to do the thinking and plan and decide and make decisions. You're orchestrating this.

Hello everybody and welcome to a new episode of Honest UX Talks. This is a very special episode because this time I'm not joined by Anfisa, my regular co-host. I'm doing a solo episode which is kind of weird since we're all about conversation but we decided to explore this format out as well and

I am going to talk about a topic that I've been reflecting a lot on for the past couple of years. And the topic is called how to become an AI designer. The first thing we will have to address is, is there such thing as an AI designer? For now it is, but...

Let's explore and unpack whether it's something that will stick around or what are my predictions around that? And how can you transition into working for AI problems? And yeah, just being part of this design and technological revolution.

We will go deeper into that in my monologue, if you want. This isn't scripted either. It's very much freestyle. It's interesting for me to see how a conversation between two people would translate into a conversation with myself without overly planning it and scripting it.

Before I start, I want to take a moment to thank our sponsor, Wix Studio, which we're very proudly partnered with. Since today we're talking about AI and AI impact on design careers and the way we work and the jobs we want, I just want to quickly mention that Wix Studio has a bunch of very interesting and useful built-in AI tools that are

are meant to support you in speeding up design processes and just covering for the parts that you don't really enjoy doing like responsiveness and wik studio offers responsive ai and this helps you with a very frustrating task of working on your desktop design and then

spending hours to perfect every pixel and then it works amazing and then you have to start all over for the rest of the breakpoints tablet mobile and so with AI it doesn't have to be like this with responsive AI in Wix Studio you can adjust your designs with a click and

with a few seconds instead of many hours, you have an entire section optimized for each breakpoint. And from there, you just need to tweak it, of course, but you've saved hours of work. So yeah, there are definitely good parts to the AI revolution. So I'm inviting you all to explore the AI tools that web design platforms like Wix Studio are experimenting with. And so with that being said, let's jump to the topic of the day. So I've done the exercise of asking GPT, since this is a very AI-esque

episode, I've asked GPT to ask me a couple of questions just to provide me with some sort of starting point for what I'm about to walk you all through. And GPT says I should start with talking about how I got into AI design. And this is typically what we do on Honest UX Talks. So I would say yes, let's start with that.

Personally, I have been lucky enough after 10 years working in the fintech industry, I moved into a role with UiPath. And UiPath was doing AI as a company. It's a robotic process automation company. So they're all about computer vision, building their own custom models, UI automation, and so on. So I was in the right spot.

space in the right company. After 4 years working on one of their core products, Orchestrator, I was invited to join the Clipboard AI team. This was the first LLM based product that UiPath built and so I joined that team.

There was some design work done before that, but there was also a lot of ambiguity and a lot of questions and they were using GPT before chat GPT was launched. I was spending a lot of time on OpenAI's playground trying to understand what is this and what

and what can it do for us and how does it really work and so yeah just bear in mind that chat GPT was not even launched at that point but I was impressed with where technology is at even just by exploring the playground that OpenAI used first to serve as these

models. I think it was GPT 3.5, if I'm not mistaken, when I joined. So yeah, it was my first deep experience or intensive experience working, dealing with AI problems, things such as how do you design for uncertainty? How do you design when you don't know, you can't predict what will happen when somebody interacts with AI? How do you deal with this non-deterministic aspect of AI in design interactions? So a lot of new questions opened up.

It was very intense. I felt like I'm mostly going through the dark, like there wasn't a lot of information available online. Now everybody kind of talks about AI and there are a lot of posts about it on social media and we're all inundated and it's this noise. But at that point, again, I think the AI content revolution and explosion happened once ChatGPT was launched and everybody kind of got to touch this technology. So yeah.

they started having opinions about it. It was the most important step in making generative AI available to people after Midjourney and DALL-E and so on. So now you can generate text and have conversations with AI and LLMs are now available, democratized if you want. Everyone can interact with them and get value from them. And so now that everybody was kind of experimenting with the space, of course, content emerged rapidly. But

When I began working on Clipboard AI, there wasn't much to read through. This was partly one of the reasons that I decided to build a course when Interaction Design Foundation invited me to create the AI for Designers course. I was a bit reluctant at first.

because I thought wait AI is changing so quickly like this year is fundamentally different from what next year will be like and then in three years nobody knows what will happen so how can I create something that will probably not age very well or

will not be relevant in a couple of years. And so with the Interaction Design Foundation team, we found a system in which the lessons, the tools, examples, the scenarios representing the exercises, the people taking the course are being given. They're in a way evergreen, even if tools change.

the mechanisms remain, the philosophy remains, the frameworks for how to work with AI will not transform very much beyond this point as far as we can see. At that point, I realized that there's not a lot of knowledge made public in the space of design and I decided to, yes, accept it,

that invitation and create this course with them. So I started my journey on Clipboard AI. It just so happened that Clipboard AI won Times Magazine Best Invention of 2023 award, which got me more visibility. And so I became kind of this, I'd say, yeah, visible designer in the AI space. I niched my content even further. I launched a newsletter called AI Goodies. And I'm also talking about AI very much in my experience in the AI space, very much in it. And

And then I joined Miro as their first designer in the AI team. There was no previous dedicated designer working on AI problems. I was the first one and I was very happy to join Miro. It was one of my dream jobs ever. I wanted to join Miro ever since it was real-time board. So I was very happy to not just join Miro,

but joined it in the AI team, which I was so passionate about. It was a very important moment. And I spent one year with them. I learned a lot about the AI space. And then I realized that I need to learn more and expand beyond the single problem space, be it collaboration or

transferring data between systems, which was the premise of Clipboard AI. And now I'm doing consultancy. I launched my own studio. It's called AIR Studio and it's from AI Revolution. Yes, I'm now dealing with more industries, more problems, more problem spaces. I'm learning even more than in the past. So I'm very lucky in that.

Now, with my intro here, let's see what it means to be an AI designer, what exactly is an AI designer and why should you even want to be one. And my take on this is that it's just something that's transient. It's an intermediary stage between this thing moving us from, let's say, building conventional interfaces

conventional design, traditional interactions to building AI interactions, but we've got to call it somehow. So in a way, the market needs to differentiate between people who have experience in the AI space by now and people who don't yet have experience. So now we're hiring for AI roles. Somebody is or is not on the AI team, works as an AI designer. But my belief is that this is just transient until we all integrate this new technology in our practice.

In our craft, in a couple of years, I don't think we will need to say AI designer or just regular designer, conventional designer. Everyone will be an AI designer in some ways. Even if you don't work on products that have AI, even if you don't work directly with AI interactions and surfaces, maybe you use AI in your work and maybe you're very passionate about AI and that would make your skills transferable to roles that need to solve AI problems.

So my prediction is that this role will eventually disappear because all designers will need to kind of touch on parts of AI, work with AI in some capacity or

even in my role at Miro, everybody in that company was experimenting with ways in which AI could come in and help them solve the pain points or the problems they had. So everybody was in a way a part-time AI designer because they were thinking about the problem space and the ways in which the problems they had to solve could be supported by AI. So we will all have to think about these

things, right? We will have to think about how AI is just a new tool in our toolbox. It's a new technology that we can use to solve the problems that people have. So we will continue to solve problems and AI will be a part of that solution. And some people will have more experience working with these kinds of solutions. Yes, sure. And for a while, I think we will call them AI designers. But then every designer will work with AI. So I don't think there'll be such a call for differentiating

a need for differentiating between yes conventional and AI design it's just we're designing solutions right and some of them contain AI the future demand of this role right so yes AI design is something that is really interesting at this moment not even myself I've been working in the AI space for a couple of years now but I don't consider myself an expert because it's a space that stills

brings a lot of ambiguity. We're still learning. There aren't very clear frameworks, practices, principles, right? So apart from the very large principles, like let the user feel in control, provide safety mechanisms, be ethical, encourage people to learn and to feel safe. So some things are...

in a way, evergreen. They're encompassing. They're large. But when it comes to an interaction level, we still don't know whether we're going to see co-pilots exist in five years or everything will be just invisible AI, right? So AI won't even be signaled as AI anymore. We won't need to retire the Sparks or we don't know what's going to happen. So it's hard to predict the evolution of this role beyond my own prediction that we're not going to call it AI.

design anymore it's just gonna be a designer that works on a problem and of course uses some AI to solve it but if you still want to explore so I'm not saying that oh don't bother with the AI thing because you will have to deal with it anyway I'm really encouraging everyone in a way I hate

all the AI conversation. It's very noisy, it's the same thing over and over again, it's overwhelming. I feel there are a lot of things that you shouldn't spend too much time on, like how to prompt, right? I mean, of course, prompting is very important.

the better you prompt, the better the quality of whatever it is you're trying to get out of the AI system will be. But at the same time, we're all going to get better at prompting, the systems will get better at understanding how to capture our intent faster than through just having to put the

load on us, the effort on us, right? So with the articulation barrier and people not knowing how to prompt. So that's an interesting skill to learn, but I wouldn't focus my career around learning how to prompt. However, if you want to prepare for the AI space, if you want to go deeper into what this is and what will be, I think you just need to start as simple as just experimenting with these tools. So experimenting with, I don't know, building apps using Cursor and Cloud and

a combination of them, right? Or experimenting with Replit or experimenting with Lovable or all these tools that are now kind of empowering us to build with AI or to improve our processes, empower us to have skills that we didn't have in the past, like coding, like just experiment with these technologies. Be open. In my newsletter, I share the monthly learning exercise. There's this new format to my newsletter and

And in one of the shared tools, the most interesting articles that you have to read. And there's like the monthly experiment that you should be exploring. And this month it's going to be explore lovable, which helps you actually build things with AI, even if you don't have any coding knowledge.

I think this is part of our how to prepare a plan, experiment with these technologies as much as you can, because it's really interesting. And I think it's not just fascinating, but it also shows you the limitations, the possibilities. It kind of gives you an understanding of that universe and you can adapt it or connect it to who you are as a designer. Like the things maybe you weren't very good at. Like, I don't know.

know maybe you don't enjoy ui design now there is a possibility in which you in a way externalize ui design and you just do the thinking you do the envisioning but then ai brings it to life right so that will also give you a flavor of who you will become in the future among these technologies and possibilities and so experiment with the tools and

if you want to go deeper, build an app, launch it on App Store. AI is now supporting you. In my last newsletter, I shared a very interesting article by a design leader from Adobe who on a weekend built a product and launched it to App Store and

everything with the help of AI. So you can do these kinds of experiments now and you can learn a lot from them. And this is the way you prepare. Just play around with them. And of course, there's some relevant literature around it. Literature is a very generous word. There are some articles and some resources like Google Pear. You should try reading through their principles and their philosophy and their frameworks and so on to understand the ethics behind AI and how we should think

about AI when we're making design decisions around it. And of course I want to recommend my course. It also has a component around how to think critically and not trust AI blindly and not just fake enthusiasm around these technologies, but actually understand that there are just things that we can now do and explore and experiment with.

And we should do it with a lot of critical thinking and responsibility and, yeah, just ethics in place. And how do you build and reinforce those ethics in the AI systems we're currently building? So explore the space, read. I think...

When I was supporting people as a mentor in transitioning into the AI space, my first or my biggest piece of advice was, hey, look, just do a project. Do your first project. Read about the theory, map out the most important parts or the key aspects of the design role, and then just go out there and start solving problems. And now for AI, it's the same. Just choose a problem and try to solve it with AI. It's the same process for learning, right? Just getting your hands dirty and experimenting.

Yeah, GPT now asks me: "How can designers practically start incorporating AI into their current roles?" The outline GPT thought about is like steps for identifying AI opportunities in design projects. And this is something that I do in a workshop that's called "Identifying AI Opportunities in Design Projects". And I'm going to give this workshop in Berlin in April and probably online beyond that, let's see.

This is exactly what I do in the workshop. So we start by mapping out a problem space, a set of hypotheses. We come together as a workshop group. And depending on the context, that might be your problems from your company or fictional problems or whatnot. Most of the times it's best to work on real problems as much as possible, right? So you start from a problem, you write down everything you know about it, you construct or surface, if you have it somewhere, just bring in the customer journey map and the pain points associated to it. And

This is again, now I'm just describing my workshop and I'm not promoting it because I'm not doing it anywhere. The conference workshop is sold out, but I might do it in the future online. So, but in the workshop, after we map the problem in detail and I give you advice around what's a good problem, what you need to capture when you describe it,

and so on, like what are do's and don'ts of articulating the problem and the customer journey map, then we move into what are the AI capabilities? What can AI do? And then after we look at all the things that AI can do today, we start making connections between, okay, so AI can do 200 things, and my pain points can be further broke down into, I don't know, a couple of

points or a couple of potential solutions. And so now we match the pain points and their potential solutions with the possibilities that AI presents us with. And so start from problems. This is how you identify opportunities, map out problems, and then look at what AI can now do for us and see if potential solutions to those problems are brought on.

by the AI possibilities, right? So AI can generate things, AI can summarize things, AI can spot trends, AI can emulate a conversation, AI can create images, right? So AI can do a lot of things. And it's like in my workshop, it's a list of 200 things and we spend a long time just going through the list, but you get the

point, right? So start from pain points, problems, everything you know about potential solutions, the solutions you currently have, conventional solutions, and then can those be enhanced or replaced by AI solutions based on what AI can do. And then what I'll

also doing all my AI talks is I give the slide where I re-imagine the design process with the help of AI. So supported with AI at every stage of that journey. So if you're doing research, you can use tools like Otter.ai to record your interviews.

or you want to just process all the information that was surfaced, right? So research synthesis. You can go to Miro and use the AI capabilities in Miro to take you from a bunch of unstructured data to more structured formats. Or Dovetail has really interesting AI capabilities as well. And then you might want to move to GPT to brainstorm or ideate or just see competitive market research now that

OpenAI also offers deep research on the internet within ChatGPT. So at every stage, not to mention the UI part, right? So you can generate mock-ups with the help of AI, you can generate wireframes, you can come up with information architecture and certain tools using AI, prompting it or training it if you want.

on your problem space so there are a lot of possibilities to incorporate ai in your design process now the key here or the way you should look at it is that you're still in the driver's seat you're at the center of this process you design it and you orchestrate it and you decide what are the parts where you need to do the thinking you are the one who brings intentionality to the process

You are the one who decides what comes next, what is the outcome that you expect at every stage when you have enough confidence to move to the next stage. So you're still in control. Even if AI is replacing some parts of the process or supporting you in some parts of the process, you still have to do the thinking and plan and decide and make decisions. And so you're orchestrating this. You are the...

in this process. And yes, AI can come in and help you as let's say your junior assistant, always available in different parts of this process. So this is another way by which you could integrate AI in your work. Just use it as a support in the way you solve problems.

Another note made by Chad GPD is how do you propose AI-driven projects to stakeholders? And this is another conversation in itself, right? So how do you champion AI? And should you even champion AI? I know that when I was working at Miro, I ended up having a lot of conversations around, do we really need AI here? Is it valuable? Does it make any sense? I mean, yeah, we're the AI team. We should do things with AI, but what if things can be done better in a conventional way? And I think that this

proposing of AI solutions to stakeholders is interesting. Like maybe you want to champion AI, but maybe some things are left unaided. So this is another interesting conversation that you can perpetually have with yourself, right? Is this a good problem to solve with AI or conventional is still better for this problem? And if this is a good problem to solve with AI, how do I convince others about that?

What are the fears that they have? So in the past, we had to fight for research buy-in, right? Research budget. So we would go and tell stakeholders, look, I know you're worried that this will be expensive or this will take much longer than not doing research. So we should skip it altogether. But here are the risks. We can make poor decisions. We can make uninformed decisions. We can go on the wrong path and not even

know and then it's going to be more expensive. All the waste that we accumulate is worse than just doing proper research. And so we would have these conversations. And with AI, I think, again, we should start from what's holding people back, what they're scared of. Maybe they're scared of infrastructure costs. Maybe, I don't know, implementing an AI solution will be expensive. It's going to consume a lot of tokens or I don't know. Maybe they're just scared

that they're going to have privacy or legal problems or compliance problems or so. There are a lot of things to be taken into account case by case. So just try to bullet point the issues that people might have when it comes to incorporating AI in their work, in their company, in their teams, in their solutions and try to work with them to kind of like, let's see what could go wrong and then mitigate.

So I think we're at the end of this episode. To recap, we find ourselves in the kind of "I can't find a job because I have no experience" because "I can't find a job and I need experience" and I think we're in a similar conundrum. Many people want to jump into the AI space and they can't jump because they don't have experience in the AI space.

now i think we have a lot more possibility to just experiment on our own have our side projects with ai i'm seeing a lot of senior designers we're like principal level that are regaining their enthusiasm for design by experimenting with these technologies and seeing the new possibilities and seeing how we're all in the middle of a revolution and that's exciting like we can be part of it we can be part of it just by trying on these tools just by experimenting crafting designing it's

exploring the boundaries, exploring how they fail. And it's just really interesting to be part of this conversation. And if you think about it like that, it kind of takes off some of the pressure of, oh my God, I have to learn AI. I'm not going to be competitive if I don't know AI. I don't think that's necessarily accurate.

I think what makes us competitive is how strong we are when it comes to solving problems and our thinking skills, the critical thinking that we're capable of and the experience we have with collaborating and engaging stakeholders and crafting narratives and really unpacking problems. And so those things are still the things that make a designer good. But yeah, AI

AI can in a way bring back some of the enthusiasm, some of the design love, because now we can do more than just design, we can actually build, we can take our craft further with the help of AI. And I think if you frame it like a playground and something that now extends our possibilities, then it's less threatening and it doesn't feel like a chore. So yeah, that's what I wish everyone would do, not feel like they have to learn AI to stay competitive.

but feel like they have a new playground they have new tools to infuse some playfulness and futurism if you want in their lives it's also a cultural revolution that we will see and in some parts culture will degrade our

our experience of the internet will also degrade for some parts, like the content we will consume. Probably we're going to see a lot of fakes. We're going to see a lot of crappy AI stuff. We're already seeing it on Instagram and it's getting a lot of engagement just because people stop to look at how bad those videos are. And it signals Instagram, oh, this is interesting. I'm going to show it to more people. And then more people kind of get stuck looking at it and thinking, what is this crap?

crap and then it's just fascinating from how bad it is but I think that all of these are like this transient phenomenons and in the future we will just treat it as a normal part of our experience both using technology and in our lives and hopefully it's going to be a good part and we can contribute to making it a good part everyone can play

experiment, come up with feedback, come up with their vision of how these experiences should feel and then hopefully at some point even be able to influence these experiences. So with this optimistic note, I want to thank everyone who joined. You have some interesting AI resources in the show notes. Make sure to check

them you can give us a review if you liked my monologue or any other episode from honest ux talks we have an episode with jacob nielsen where we talk about where the ai industry is heading towards and yeah we have a lot of interesting content on this topic since this is what i'm mostly concerned with in the past couple of years and so thanks everyone and hope to see you all in past and future episodes bye