#92 AI's Impact on the Design Industry w/Jakob Nielsen
45:17 Share

It's very, very hard to retrofit simplicity onto something that is very complicated and built from an engineering perspective. So I actually fear that that same thing is going to haunt us once again, that if we have too many years of AI just growing from a tech perspective and not from the user interface and user experience perspective, that that may build some infrastructure that's going to be hard to fix in, let's say, 10 years when it will be obvious to everybody that it's going to be fixed.

This makes me think about a need in the design industry for designers to step up and start articulating some principles, just be more mindful and intentional about how we build the AI-powered products. ♪

Hello UX friends and welcome to a very, very special episode of Honest UX Talks. In an exceptional edition, today I'm joined by Jacob Nielsen. Yes, Jacob Nielsen, I'll say that again. The king of usability himself, an incredible pioneer of UX design with over 40 years of experience in the industry. It's an absolute honor for me to record this interview and have my second conversation with Jacob, this time in public.

For context, last month me and Jacob connected in a friendly one-on-one chat where we discussed some themes around AI and design. And I realized that it was so fruitful and interesting that it should have been recorded and shared with the design community. So here we are. We'll be talking about the state of AI and design right now.

If this is history repeating itself or not, if AI will eventually replace designers, which roles and parts of the process might be threatened, what we can do on an individual level to stay relevant and competitive, what we should worry about when it comes to AI, and my favorite part,

what we should get excited about. Jacob, it's a privilege to have you on our show. Thank you for being here. And I want to start by opening our conversation with a large, very open-ended question. How do you see the state of AI and design right now in a nutshell? Are there any new themes or is it history repeating itself? Have you seen this before? Actually, I think I have seen many things before. I think a lot of history is repeating itself. But

But at the same time, of course, there's also a lot of new things. And I think specifically when you talk about design, I feel like there's a very exciting thing about AI user interfaces that's different from before. So I say it's the third paradigm in design.

user interfaces. And so to really recap history, the first paradigm was batch processing, which is really no interaction. So you would write up all your command on punch cards, often give it to the computer that was in the basement somewhere. All those commands would be run. And if there was no errors, then you would get a printout with the result. If there was an error, it was all for nothing.

So then the second paradigm was command-based interaction. So that is anything from line mode, so DOS, Unix, to graphical user interfaces, so Mac and Windows, where you click on the commands rather than type the commands. But really, conceptually, it's the same. So the user will issue a command to the computer. The computer will run that command, present the incremental result.

And if it's good, you proceed. If it's bad, you can undo or modify your direction. So you kind of meander or wander your way through the user interface towards your goal. And so commands.

But AI flips this, and this is what I call intent-based interaction. So rather than saying, do this, do A, do B, do C, which is command-based, you say, I want this. I want a picture of a cow in a beautiful landscape or whatever it is you want. And the computer will produce that. The AI will produce that. So intent-based, you say what you want. You don't say how it should be done.

And so if you think about that middle paradigm, but most of us have only known that really, but command-based. So it has in some sense, some advantages, which is computers are very kind of obedient. They'll do exactly as they're told, but it's also a downside because they will do exactly, exactly as they're told.

So if you say something a little bit wrong, right, the computer will just do that. It does what it's told. And so the benefit of the intent-based, it's more interpretation-oriented. Now, obviously, this also means interpretation can be wrong. And that's one of the weaknesses of the current AI. And we hope it'll be better in the future. But I think that's a big, big, big change. I mean, that's actually a true revolution. First time in about 60 years that we have a completely different way of interacting with computers. So I feel that's a big change.

Now, okay, so you also said, is history repeating itself? And I think, yes, because for all of these previous generations of user interfaces, when we got a new way of doing user interfaces, and this can also be a smaller modification, like, for example, going from local PCs to websites, and websites are just graphical user interfaces, but it's still kind of different technology, or going from websites and desktop computers to

to mobile interaction. That's again, a small change, just a small screen over a big screen, but still a graphical user interface. Well, for each of those changes, you know, usability would take a big hit just in the beginning because every time there's a new technology, what happens is that it's driven not by user experience people who know what people need,

It's driven by engineers and technologists and programmers who knows what computers need. Because in the beginning, logically speaking, it's hard to make it happen. So that's where all the emphasis is not on what people need. And that's why you see a lot of designs and products in the really beginning of any new technology that actually has horrible usability.

They do not do anything that people need really. And so they often fail products as well, but there's also excitement about them because they show new possibilities. And that's also with AI. It shows a lot of new possibilities, as I just mentioned, but the actual usability user interfaces are often very bad, like chat GPT, enormously long scrolling, mid journey, very obscure commands and parameters.

It's really terrible. But then you get like beautiful pictures for free. And so that's why people take it anyway. But honestly, anybody who knows just a little bit about usability can look at those user interfaces and say they could be so much better. And that's history repeating itself.

It's my first tech revolution, so I can't really trace it back to anything similar that I've experienced in my career so far. In my, let's say, 10 years in design, this is the first time I'm part of something this big. I wasn't there when the internet emerged. I wasn't there when the mobile phone revolution happened. So for me, everything feels new and probably a bit scarier than for someone who, like you with experience, has been through these revolutions before. I just want to add to that.

that you're right, it's very exciting and scary at the same time, every time this type of thing happens. But I also think that the history experience shows that things most likely will get better. And also that in the beginning, actually, to be honest, we should also point the finger at ourselves. I mean, the UX people, we have some of the blame as well, because what happens typically is that the UX people

are very busy and very engrossed in doing old school design. Like they used to do PC design when the web came out and then there was advertising people who did the website, not the user interface people. Or when mobile came out, people said, oh, we'll just take a website and shrink it to be small. Well, shrinking it to be small is the

wrong way of doing a user interface for a small screen, and so forth. And actually, when AI started to come big, which I would think it was probably in 2023, very few UX people really worked on it. I mean, most people say, "Oh, that's some hype or whatever." And I really don't think it's hype. I think it's a very, very big thing. Maybe we can discuss that as well. But it's also our own fault. It's our own fault for kind of like neglecting these new technologies

and focusing on designing for the old technology. Because that's a big thing. I mean, financially, the old thing is the biggest in the beginning. The new thing is small in the beginning. But again, history repeats itself and it's growing, growing, growing. But if it grows from a bad design base, that comes back to bite us years later. And I like to refer to Microsoft as an example because the Microsoft, their main products, office products, you know, Word, Excel, PowerPoint,

they were exactly built on this old school notion of programmers like let's add, just cram as many features as we can into the software. And they did not have any good designers and usability people, at least maybe they had one or two, but they didn't have enough for those big product lines. And today they do. Today they have a big effort. But even, I would say, 30 years later, they are still suffering under these products at the core, having been designed the wrong way. It's very, very hard to retrofit simplicity completely.

onto something that is very complicated and built from an engineering perspective. So I actually fear that that same thing is going to haunt us once again, that if we have too many years of AI just growing from a tech perspective and not from the user interface and user experience perspective, that that may build some infrastructure that's going to be hard to fix in, let's say, 10 years when it will be obvious to everybody that it's going to be fixed.

This makes me think about a need in the design industry for designers and the entire industry to step up and start articulating some principles, methods, solutions, just be more mindful and intentional about how we build the AI-powered products we

work on or just a simple act of being more vocal around how an AI product should ideally as much as possible behave. I think it's an important contribution. So even just as a public discourse, we should be talking more about this. While you were sharing your observations, I was thinking, why could it be that designers aren't extremely excited to work for a completely new paradigm?

for a completely new philosophy and new problems. Like there's so much unknown to unpack and discover and understand and design for. So in a way it should be very exciting, but I also feel that there's a lot of inertia. It's also in itself, all this unknown poses a huge challenge. Like sometimes I work as a product designer for AI problems, right? So sometimes there's nothing I can refer back to. There are no mental models. There are no standards. There are no answers.

I think it's also discouraging a bit for the design industry. But yeah, I hope we will mobilize sooner. And I also hope that tech companies and tech in general will understand that there's immense value in embedding design early in this revolution and not having, like you said, retrofit and just course correct and fix bad decisions, design ignorant decisions that they're making right now. So

You also touched briefly on the question whether AI is a hype or not. And I'm interested if you could expand your thinking on that. Is this just a wave of excitement that's going to pass quickly like we saw recently with NFTs and crypto? Or is this something else that's going to stick around for a while?

I think that it is not hype. And the reason for that is that there's a lot of research now that shows very large productivity gains from people using AI for a lot of different work. And so that to me says that it's real. Now, there's a lot of other technical innovations that sound good and you can make up a story about how it would be wonderful if we had it. But those are stories. And we don't actually know if for a fact it's going to be good or bad.

But for AI, we actually know it's good. We know that it increases productivity. And of course, depending on which study it is, the number will be different. But it's usually around 40%, 4-0. That's enormous. I mean, I've done a lot in my days of kind of quantitative usability studies trying to measure how much better one product is than another. And they are usually much smaller percentages. And so if you can gain 40% more productivity

productivity, that's huge. And that's what the current AI products and I just mentioned that they have bad usability. And also technologically speaking, the actual AI is also still only kind of sort of some of the early products. And we know that they get better at a relatively fast

pace. And so you would imagine that next year, the products will be better. Hopefully, the user interfaces, usability will be better as well. And that would then mean even more productivity enhancements. So to me, that shows it's not hype. I mean, if you can go to basically any business manager in the world and say, I have this, and it's going to make all your employees 40% more productive, that person is going to say, where can I sign in the dotted line? Particularly since they're relatively cheap products, actually. So people will want that. And

And then now comes into what may be hype, the speculation. So the speculation would be that these measurements of 40% productivity come from people doing the current work, but with AI. So some of the studies that have been done are things like business professionals writing proposals or management consultants doing a strategic analysis of some market area or law students writing legal briefs, but they do it much faster and actually better quality as well.

speed and quality both. So there's not a trade-off. You actually get both improvements. So those are great. But what I predict, but I don't know, right? So this is where it may be hype, is that we should also restructure how we do work. And that I think can lead to vastly more improvements as well. Not just do the same work

faster, but do different work. And that is something that's much harder to invent and come up with. And that's something where, you know, you talk about organizational inertia and changing big organizations and big bureaucracies. That tends to take 10, 20 years. So what I think is happening is very soon, some people already know

Other places, maybe over the next one or two years, they're going to embrace these current type of products and do their current work faster and better, and they'll get a big productivity increase. And that's why the products will sell and already are selling a lot, actually. So the next stage, I think, is a bigger change, but that's only speculations.

So that may or may not happen. I think it will happen, but that we don't know. But what we do know is on an individual basis, current work, you can get vast improvements from taking current AI, not the future, but the current ones that are already available and the work you're already doing, just do it faster, do it better. And that is not hype. That is true economic value. So therefore, it will happen.

And just a small thought I want to add there, it's really interesting to think that AI has actually been a part of our lives for a long time now. We're just becoming aware that these technologies are penetrating our everyday lives and becoming household behavior, if you want.

Waze has AI, or Spotify uses AI, Netflix uses AI to personalize content recommendations. And so it's here to stay, clearly. I think that we're going to see it grow into more aspects of our lives. And this is where I think that we should perpetually call for responsibility when it comes to making design decisions. It's

essentially going to change not just the way we work, but the way we live. So it's changing the human experience. And so it's interesting to think about it that in the present, we're making decisions that essentially shape the future of how we live. On this note, I want to jump into probably the juiciest or most stressful question in the design industry right now. Will AI replace designers? What are your thoughts about that?

I actually don't think so, because I think that the best performance comes from combining AI and humans in a synergy effect. And so they both have their role. I actually do not believe that some people are saying, well, there are all these things that only humans can do, like empathy, understanding users and so forth. But that said, I also do think that

There's a lot of things that benefit from human involvement and AI will not do as well. And so if you combine the two, so AI has a certain type of creativity, which is the ability to come up with a lot of very different suggestions very fast, which if you have humans, you've got to get 10 people in a room for a day. And that's very expensive for AI to say, give me 10 proposals, bang, and it's 10 seconds later, you have them. And of course, many of those are bad, but that's the same with human brainstorming as well.

Then you have to have that discernment of the ability to select out of those 10 or 20 or 50 proposals from the AI, which are the ones you're going to pursue and refine. And I think there's a lot of also doing usability studies, for example. I don't believe in this notion that you can have kind of artificial users, like you can just ask the AI which design is better, A or B. One of the reasons we actually do usability studies in the first place is that humans are actually quite unpredictable and have

very odd behaviors that we would never have thought of. And, you know, one of my kind of sayings about user research is that if you do a user study and you didn't discover something unexpected, it shows you ran the study wrong because there's always something unexpected.

If you're just going to confirm your existing ideas, you didn't do a deep enough study. So there's always something new that comes up. And so you've got to see real users use your products and things like that. And then you combine that human insight with the AI ideation, ability to write things, ability to translate things into multiple languages, a lot of capabilities. And this is actually only what we have now. So I think that in the future, we will have a lot of it.

very interesting design capabilities from AI that we don't have so much now. But for example, the ability to present information at different levels of complexity. And I'm really talking mostly about written information, but the same could actually be true for

or other media forms as well, because AI is getting to be very multimedia. But in text, it's kind of easiest to understand it because in text, we know that people have very different reading skills. Some people are highly educated, they're overeducated academics, and they read very complicated, long sentences. And other people hardly were able to get out of school and they

can only barely, barely read the minimum. And you can't show the same text to both those types of audiences. So you want to have the text written at a level that is appropriate for the audience. And some of that you can do manually. You can write two or three different versions, like very simple, somewhat medium, and very advanced. But to really, really make it suitable for the individual user, each person is a unique individual after all. That's what AI can do

It's not necessarily perfect at it right now, but I think that's one of the things we will be getting. And a great example of this is some hospitals actually starting to experiment with doing patient information this way. So they have what's called, I think, an after-visit summary. Like when you've been to see the doctor, they will tell you, well, this is what we discussed and this was the diagnosis and this is what you should do to get better.

And particularly what you should do to get better is very important that people understand that. And again, if that's written in a very convoluted way that the way a physician might naturally write it, a lot of patients can't understand it. And so you can have the computer or the AI rewrite this information to be,

understandable at that patient's reading level, which again is very individual. And some people understand certain things, other people understand other things, and maybe also they have a language issue, and maybe they prefer a different language than the language of the doctor, and so forth. So many, many, many reasons why you may want to let the patient read something different than what the doctor wrote. And AI can do that individualization.

of the medical information so people have a better chance of understanding what's the recommendation to do to get better or what medications they should take and should they take in the morning or the evening or whatever, 50 other things that can become very complicated. So I think that's an example of where AI can really enhance and save lives, literally, but it just enhance our experience, I would say. So that's kind of more the bottom line there.

This is a reason to get excited about AI, right? So there's actually a very tangible possibility and it's happening now. I think the medical industry is being revolutionized by AI at a very impressive pace. It's making the biggest impact so far. Thinking about the areas which AI can be a very fruitful co-partner or let's say co-creator for designers. You've mentioned brainstorming. It has trouble, of course, replacing research efforts. And I think it should

remain like that for a long time. But are there any specific parts of the design process or maybe design roles that you feel are most threatened by AI? My question comes from a place where I'm thinking that UI design could be very easily automated with AI today. And what will happen to the UI designers that are now designing these pixel-perfect Figma files if AI can also do that with the right prompt?

No, I think it can. I think a lot of those maybe what you might call lower level production style design can be automated and should be automated. I mean, if things can be done better, both in terms of speed and cost and quality, well, then you should do it that way. And then those people should do other things, whether you take more advantage of their human skills. But that means that they have to retrain and they have to think of themselves in different ways. And some of those skills that were highly prized in the past,

do not necessarily become highly priced in the future. And I mean, certainly not necessarily so much in user interface design, but other types of graphic design, like the ability to draw nice illustrations. I mean, you can get very nice, pretty pictures out of Midjourney now, and maybe not like the world's best illustrator, you know, but you can certainly get it at the quality of a lot of the people who work on freelancers and things like Upwork, Fiverr, whatever those different services are called.

A lot of that, I think, will scale back because it can be done by AI. I said better, cheaper, faster, more variety. All those qualities will be available. And so therefore, that will not be done by humans. But I don't think that means that you will not have design. It just means that there will be some other types of design. And also some types of design that we don't even know of today. Like I mentioned before, this idea of being able to rewrite information for individual patients is

Well, there still needs to be some design behind that, but the design is not the actual writing of that message that's done by the AI, but it's going to be kind of like the rules or the guidelines for how that's done. And that requires insight and usability and insight in this example, medicine, and in writing and content strategy and all those type of things, and then coming up with some rules.

In this example, the writer's job is not actually to write the actual copywriting, but to rather present tone of voice criteria, other criteria for when you do various things. And I don't even know what those questions are because it's kind of new, but that's something we have to develop and discover and find out. And that's why, going back to what you said before, I think it's scary for a lot of people because they will be asked to do things they have no idea how to do because nobody has any idea how to

do it. We've got to make it up. But is history repeating itself? Because in the very beginning of UX, I'll always trace it back to the design of the push-button telephone at Bell Labs in the 1950s. They had no idea how to even do user testing on a telephone. You know, that hadn't been done before. Telephones were only designed by engineers. And now the actual user interface was push-button, so they had to design it and they had to find out how to test it. And they discovered that, you know, they've discovered that. And now we don't

They said, do our usability research the exact same way as they did in 1950. But the ideas actually have remained. Test with real users being the most important. But somebody had to come up with that point. And then all the other things we have learned since then, when the web came out, people thought, oh, it's just like you put a brochure and you can download it over a modem. And that's not a website. A website is not a brochure. We discovered that pretty quickly. But in the beginning, that's what people thought.

thought. So we have to discover these things and we have to learn them. And yeah, you're right. In the very beginning, each individual person has got to learn it from themselves. And that's tough. That's really hard work. Later on, we're going to have like some best practices. And then once something has been around for a long time, like web design, now we really have extremely detailed guidelines. We really know how to make a good website. We don't necessarily know everything, but we know a lot. So being a web designer today is not necessarily as

revolutionary as it was back in 1996 or something like that. It was just brand, brand new. So that's as it become easier, right? And then you move to this new field and it becomes, ooh, scary again and new again, and we don't know what we're doing again. And that also means that, yeah, we'll have to accept we're going to do some things that turn out to be wrong because you've got to push an experiment. But see, this is where the beauty comes in.

of user experience because we have a methodology to find out if we are right or wrong, and it's called usability testing and other types of research. So we have a correction mechanism. So if we only bother to do the research, and I know many people don't, but if you bother to do the research, then you might attempt something that turned out to be wrong

And that's no shame because it's new. So some percentage of what anybody does will turn out to be wrong, but we can test. And when we find out it's wrong, then we don't ship that. We don't inflict it on the customers. We go back and say, oops, that was wrong. Got to do something else. And you try another thing. And you try enough things, you find something that works, and then you ship that. And I think it's actually quite exciting.

to be part of the reshaping of our roles. We have the opportunity of redefining how the design role looks like. If UI will be automated, then what's left for us? Or where are the parts where we can bring more value? Or if we want to retrain professionally, what's the expertise area? And maybe there's no clear answer right now. Most probably there's no clear answer. We have to answer that by understanding what's missing in the AI and design industry. At that intersection, how can we contribute? The

The point I'm trying to make is instead of feeling scared and overwhelmed and worried, I think we should feel excited that, okay, things are changing. Change is not easy. There's also this natural human resistance to change and the worry that maybe I will not be relevant in the future. But I think that we can individually go on a pursuit of answering the question, how can I be relevant?

relevant? How can I contribute to these standards and to the way we shape the rules, right? You were talking about web design and how things are very clear right now in terms of how a website should behave. We now have the opportunity to stop pushing pixels if we feel that's not going to be around for a long time and move into articulating bigger principles, solving bigger problems, and just, yeah, using the critical thinking skills that will probably not be replaced by AI anytime soon.

Critical thinking is one thing that I think we should all exercise in this transition. But my next question is, in building on top of this, what's your advice for a designer that wants to remain competitive in the future? What can they do to make sure they can still work in design in the upcoming years?

Well, I mean, I think it's going to be a gradual change. We talk about this revolution, but in real life, it becomes more of a slower change. And so because you cannot change the entire organization and the entire way everything is done in one day or in one week. I mean, that's the thing that takes years or probably even decades. And the same is true for humans as well. I mean, we all have to learn a lot of new things, change the way we're thinking, as you just mentioned. And it

And again, you can't quite do that just overnight. But what I do think we can do overnight is we can start pushing ourselves. And particularly when you talk about AI, you start pushing yourself by using AI for more and more things, including things that are a little bit

where the current technology may not actually be perfect and where you have to like roll it back and say, well, those ideas didn't quite work anyway. But if you don't push the boundaries, then you're also not going to conquer new territories. So you've got to actually try additional things. Yeah, for anybody who's listening who has not really used AI yet, they've only seen demos and stuff, there's a big

big, big difference between seeing a demo and doing things for yourself. And so that is absolutely the step one. It's just even start using some of the simple AI tools that are available on the public internet. Anybody can subscribe for $20 a month to chat GPT, to mid journey for artwork. I mean, those types of services. Then there are the more internal, more advanced tools and more specialized tools. That's where I actually think the big value is.

not these generalized tools. But that said, the generalized tools you can start with because they're cheap and they're easy and anybody can use them. I mean, even if you just create things like not a professional design, but like a birthday card invitation or something like that, you can do that in mid-June. You can ask ChatDPT to...

help you write for the next report, you know, like write 10 different heading headlines for this section, and you pick the one that's the best. I mean, there's a lot of things, or come up with different design ideas for doing something or other, or give me a proposal for a research plan. I want to find out the following things, what could be ways of

of discovering that. And again, it's not going to write you the perfect research plan, but it saves you a lot of time. It's again, it goes back to that productivity game that it saves you a lot of time from some of the mechanics of just getting down on paper or into your word processor. Some of these steps it'll just give you, and then you've got to edit it and change it and oh, completely overlook this point that's very important for our industry.

And then you put that in yourself. But you can get a lot, but you can only get it if you do it. And so then gradually you'll get better and you'll discover more advanced or intricate uses. And then you can graduate to the more specialized tools of which there are many. I just think, honestly, it's like just get going in step one and then keep pushing in step two. Those would be sort of my two pieces of recommendations.

I'm going to build on your advice and follow up with a question. I'm following you closely on LinkedIn. I follow your activity on uxtigers.com as well. I know you're playing with AI tools, experimenting quite a lot individually. What are the tools that you're excited about right now? And what are the tools that a designer that, let's say, hasn't been exposed to AI yet should start with?

Well, I think, honestly, if you want to start, you should start with some of the more simple tools, like start generating pictures in Midjourney or start generating text in ChatDPT. But what I'm actually very excited about is there is this multimedia push where you can make songs and music, you can make animations, you can make actually videos, not up to the true performance.

You're not going to think that these animations are like what Walt Disney Studio has spent years of making a movie, right? But maybe in a few years you will. And then the distinguishing point does not become that mechanics of drawing hundreds of thousands of frames. Actually, they stopped hand drawing many years ago already, but that's in the old days. If you look at something like Snow White, which was the first big, I think, Disney movie,

People were sitting by hand and drawing hundreds of thousands of drawings to make that animation. And now, as they said, it's done by computer. But now it will be done even more so by computer. You will have less illustration talent, but more storytelling talent. More sort of like, what is the point about this? What are we trying to communicate?

And I think we already have evidence that unleashes enormous amounts of creativity in things like YouTube and TikTok in videos. And again, if you look at those videos that are on YouTube or on TikTok, you will never really confuse them for a professionally produced Hollywood film. But they get billions of views. I mean, they have watched enormously much.

And it's because they speak to people on kind of a different level. And many of them would be about things that I would never, ever want to watch. But some talk to my interests and some are interesting, exciting to me. And the same, this is multiplied by millions of people. And so every individual person has things that they actually do think, oh, funny cat videos. No, I don't really want to watch funny cat videos. Maybe I want to watch some other funny videos and so forth.

The point is these specialized things get enormous viewership. And now we are unleashing that creativity one more level. I mean, we've already unleashed it some in that you can produce a video on your mobile phone that's already huge compared to the old days where you took a studio full of lighting people and five camera crew and all of that.

but it's going to be even more so now. And so that multimedia creativity, which I think also has some downsides, but I think it has a huge upside, which is kind of what I prefer to be positive. And the upside really is that billions of people can be multimedia creators. And for example, this notion that you can make a song. So first of all, most people cannot compose music. People may be some medium good at actually writing the lyrics. Most people are bad at that.

People have very varying singing talents, to put it mildly. But, you know, you don't have to have any of those. You just have to say, I want a song about such and such. It'll make you five songs. You listen to them and say, okay, this is the one that I want to play for my wife's birthday. I wonder whatever it might be, right? But you can get a song for an extremely precise, maybe even only for a party of 10 people, and you play it once, and that's the only time that song is ever played.

That's this micro creativity that we can get now in all of these media forms and in video as well. Certainly have it for pictures now. Pictures can be anything from kind of cartoonish drawings to things that look almost like a real photograph. And that's also, by the way, one of the potential downsides of so-called fake news or deep fake or whatever people call them, that it becomes easier. I mean, it has always been possible, you know,

So it's nothing really new. But again, when something becomes easier, more of it is done. That's one of the usability of big principles. If something is made easier to do, more people will do it, more it will be done. And the happy news about that is all that specialized creativity unleashing hundreds of billions of millions of people and enormously many, many people can do very specialized things. Some of those will become big hits. Most will not.

But even the ones that are not big hits have a purpose for maybe an audience of five people, maybe an audience of 100 people, and then we'll be fine because they will cost almost nothing to produce. And then the downside is all this kind of deep fake and fake news and problems, I guess, that arise from that because they become more realistic looking and they become easier and cheaper to produce, which will mean more than will be produced. So we have to have some ways of kind of counteracting all of that.

To be honest, what I feel is almost like the biggest risk. I mean, it's much discussed about like all the political impact of you can make it look like the prime minister said something stupid or whatever. Yeah, yeah, that's also bad. But I actually think the worst is probably the completely individual spam or what's worse than spam, which is called phishing, which is kind of malware trying to get cheap people out of their passwords and things like that, because that can be done hypercritically.

hyper individualized, again, which is positive when we talk about creativity, but it's negative when we talk about criminals. And so you can have this very hyper individualized emails or even fake video that looks like you're talking to your nephew or something like that, but it's not. And that is very, very dangerous.

But that is actually one of those cases where I feel like a technological solution is required. So I'm, of course, usually not the one who says that it should be technological. I think it should be human. But for this question of basically fooling people with fake data, I feel that that's where you have to have some kind of technological solution so that

it'll just come up and say, this is not your nephew. I mean, or whatever, that the computers should have some way of doing that. I'm not the one who can invent exactly how that should be done. But I feel like the internet right now is still too much like it was back when it was invented, you know, in the 1960s or something like that, when it was just a few university people and all they would do is send research papers around. And there was no need really for security because few students will play some pranks and

send some funny things around, but there was nothing really bad happening, right? But today there absolutely is. And the inset just needs a completely different level of security than when it was invented. That's again, a great example by the way of how hard it is to retrofit. Because the inset was invented under the assumption it needed no security. People would just always be nice.

And the worst that could happen was a practical joke. That's not true today, right? But it's very hard to put that genie back in the bottle. That, I think, is really the more scary or dangerous thing, which absolutely will happen. So we just need to start working on it. But again, I want to focus more on the positive. And the positive is enormous individual creativity is going to be unleashed.

I love this idea of AI becoming a vehicle for self-expression for everyone. Like until last year when Midjourney, Dali, Stable Diffusion and GPT became so popular, some people couldn't even find the instruments, the tools to express themselves. If you are not naturally drawn towards something in particular or have studied painting or drawing or anything, how do you

express your feelings? How do you express your ideas and your life experience? And so AI offers a vehicle for that today. It's in a way democratizing creativity and self-expression, and it's amazing. I think it's also threatening for the people who make a living out of art and expressing themselves. Like everybody has something interesting to say in a way. So if

Everybody has something interesting to say. How can you still make a living or a career out of expressing interesting ideas? I also wanted to add something on the danger or cautionary tales around phishing and deepfakes and how this technology can be used for bad intentions and evil purposes.

And a good example to keep in mind is how AI in a way is contributing to the fake news problem because it can easily create fake news content and then propagate it on the internet. But at the same time, with the right guardrails and mechanisms in place, we can also flag fake news content.

on social media. And Facebook is already doing that, probably not to the extent it should, and many parts are escaping the system. But we can put mechanisms and guardrails in place with the help of AI to, funnily enough, counteract the bad things that it is doing. And so I think that this is an interesting space in which, as designers, we could spend more time in, instead of worrying about UI automation and how mock-ups will become AI-exclusive.

Here's an interesting problem. How do you fight deepfakes? How do you label them? How do you flag them? How do you put the right guardrails in place so that this technology doesn't go in a very dark place? And I want to really emphasize that this is a user experience problem. I mean, I think there's a technological side to it as well, like various ways of authenticating whether a photograph is authentic.

It's a photograph for its AI generation. So there's a technological side, but then it still ultimately is a usability problem because you've got to communicate to a person or to masses of people. But on a one-on-one basis, at any given time, one person is seeing this one photo and what do they make of it? And so that becomes a user experience design problem.

And again, sad to say our expertise does not tend to be involved. It tends to be a lot of kind of like political discussions about what they want and don't want and some engineering discussions about how can they code for that. But then that third parameter

which is how does it actually work with people? And it is a person problem at the end of the day. It's usually ignored, but that is our responsibility because like how would, like I was talking about this, whatever some random prime minister, how would that person know about UX? It's nothing you learn about in politics, right? So all these people or engineers still, I mean, some engineers know now that, you know, the people who actually work on, let's say, the type of products that have a lot of UX in like websites or certain PC applications and all, they know.

but a lot don't. And so again, it's our job to communicate that. And again, there's a whole host of additional areas where our expertise could help increase quality of life for the world population. And there's a vast amount of service design problems. And I mean, to this day,

Most user interfaces are not actually designed user interfaces, but rather a side effect of the implementation. A lot of products you could just buy, even like a toothbrush, as an example, it has a user interface that's getting very complicated and yet is very awkward because it hasn't really gone through this design process. I honestly don't think there's any risk involved.

I mean, this is a classic question is, oh, is this going to cause all this unemployment because now? Well, yes, if each person can do twice as much work, I said the current statistic is more 40% work, but I think in 10 years it's going to be twice as much. Yeah, if each person can do twice as much work,

The obvious first conclusion is we only need half the staff to do the same work. But we don't need only the same work because UX is so enormously needed for doing more things. And so that's why I don't think there's any risk of unemployment, because I feel like there is vastly more things that can be done. And again, if you think about it in sort of economic terms, if each person can do twice as much work, that means the price per work or per thing done is half of what it used to be.

Well, if the cost of UX is half of what it used to be, that means that many more companies or people will do it for many more projects because honestly, it's too expensive now. I mean, my entire career worked on trying to make UX cheaper, but it's still pretty expensive. It's very handcrafted. It's sort of very old school experience.

type of craftsmanship, and that's very expensive. And so if we can cut the cut over the next two, three years, cut the cut in half, over the next 20 years, maybe we can cut the cost to a quarter or something like that of doing UX work. If we cut the cost to a quarter, I don't think that means that four times as many projects will get done. I think it means 10 times as many projects will get done because there's a vast, vast range of areas that would benefit from UX, but they just can't afford to pay it right now.

And so if we can do our work cheaper because of AI, that does not mean we'll be unemployed. It means we'll do more different new things. And I say new exciting things, really improve the quality of life for people, which is that's our real job. Our real job is not to draw a box on the screen, but it's to make life better for people.

I love that. And I think it's the perfect idea to start wrapping up our conversation. I'm going to invite you for one last reflection. If there's one takeaway or one, let's say, mindset framing that people should walk away with from this conversation or anything else you want to mention closing this episode and this incredible conversation, what would be that last thought or idea or the main takeaway? I think my takeaway is the future is going to be beautiful.

There's so many more things we can do. I think there's an inherent human tendency to always look at the negative. And if you talk about AI, if you read, I guess, what people post about AI, there's a large number of negative things. And some of them are even true, I mean, or at least there are potential downsides that it's good to be aware of. And if you're aware of them, you can actually minimize how bad they become. You could probably not eliminate them, but you can reduce them.

But the upside is vastly bigger. And that's whether you talk about AI specifically, you talk about user experience, many other things for that matter as well. But that's not our area. And that's the other thing. I think we have our expertise. There's some ways, some things we can do to make life better for people. There are other things needed as well. We're not the only thing in the world that's important. But anyway, that's what we can do. We can make technology suitable for people, which it still is not. There's still so many awkward things happening.

in using computers or using other things that actually have become complicated because of proliferation of technology. We need to beat that back and make it easy again, make life easy. And we can. I mean, we know how to do it. Some of the new technologies, as we discussed about in this podcast, we don't know exactly how to do it, but we can find out how to do it. And so that's my kind of bottom line. The ultimate thing is don't be depressed that everything is perfect. Think about

In general, we can make it enormously much better. And this is billions of people we are talking about, rich people and poor people in developing countries where they don't have all the fancy things, but we can make it so, I mean, they can make it themselves, but we can help them make it such that quality of life goes up dramatically worldwide for billions of billions of people, individual creativity, cheaper products. I mean, whatever, lots and lots of things. It's hard to...

even mention how much it can get better, but it can really get much better. So I think don't be negative. Don't focus on the downsides. Be aware of the downsides, but it really can become so much better. And we are the ones who can make it.

We can make it so. So that's, I guess, is the conclusion. Make it so. I'm going to take away the point about how AI can actually address some of the inequalities in the world and solve the privilege gap, help more people have access to technology and understand how to use technology and enable them to navigate the hardships and have better opportunities. So this is another reason to get excited about what AI can bring.

Thank you so much for joining us, Jacob. It was incredible. I feel like it was the best conversation I've had. It's going to be really hard for 2024 to have something that's more of a highlight than this conversation. So the best part of the year has already happened for me. Thank you so much for joining Honest UX Talks. I'm very grateful we got to speak again.

And I want to invite everyone to continue to learn from you. You can follow Jacob on LinkedIn. He's active there. And also make sure to read his insights-packed articles on uxtigers.com. And with that, thank you again, Jacob. Thank you. This was a really great, great show. And thanks for leading me through all these interesting topics we discussed. Thank you. Thank you so much. Goodbye, everyone. And see you on the next episode. Bye.