Is Generative AI Killing Creativity Or Enhancing It? — With Scott Belsky

0:00Let's talk about the future of creativity in the AI age with Adobe's Chief Strategy Officer right after this. LinkedIn presents.

0:13Welcome to Big Technology Podcast, a show for cool-headed, nuanced conversation of the tech world and beyond. And today we're joined by Scott Belsky. He's the Chief Strategy Officer and Executive Vice President of Design and Emerging Products at Adobe. And today I'm so excited to have him on to talk about how AI is going to impact creativity, whether it will homogenize it, as well as a bunch of other fascinating subjects that Scott is looking at in the tech world. So glad to have you here, Scott. Welcome to the show. Hey Alex, thank you. Great to see you. You know, we've talked about this in the past, but when I was just starting out way back in the day, I think I was like working in my marketing job at the time in New York City.

0:55It must have been 2010, 2011. You gave this presentation while you were the CEO of Behance, kind of talking about the future of community and creativity. creativity and it was maybe it was in like some small office in New York City. And I remember kind of cramming in the back and being like at the end, like, who is this guy? Because clearly you're able to see this stuff in just like a very clear and interesting way. And it's led you obviously to great places in your career. So it's very cool to be able to sit down and speak with you on the pod. I appreciate that. Thanks. Great. So let's talk about creativity.

1:30Do you think we are more creative now than we were in the past. I mean, let me just set it up. It does seem like a lot of creativity is homogenized now. Everything kind of looks the same, same format, same feel, same amount of creativity. It's driven by the data. Where do you see the trajectory of creativity today? Well, I'll make two points on this. First of all, I feel like creativity is the world's greatest recycling program. Everything that we get as inputs generates stuff in our brains that then generates outputs. And so, of course, what we are exposed to influences what we put out in the world.

2:10And you do have to wonder if we're all exposed to mass media and we're all getting the algorithmic, whatever's popular and trending and mimetic and everything else, does that actually create more of a center of gravity around kind of the main narrative that maybe makes a lot of more like homogenized outputs, you know, I'd say that's a really interesting question. And, and perhaps, right. But of course, any great creative mind will tell you that they expose themselves to crazy things. They, you know, they go traveling, they, you know, they try to find the mistakes of the eye that, you know, that inspire new things.

2:50And that's, of course, where creativity comes from. So I don't I don't think that maybe there's like a there's a new a new the technology is sort of of social feeds and stuff, you know, counter some of this. But I still think that there's genuine human ingenuity that drives new areas of creativity. However, you asked this question around, are we becoming more creative? And I happen to think that, unfortunately, humanity is typically peak creative at five years old because we feel like everything we put out there gets the smile of a teacher and the refrigerator positioning from our parents. And we get this, like, we have this incredible creative confidence because we feel like everything we make and draw is amazing.

3:37And then we slowly realize that there's people with skills that we don't have. And then there's these people called critics that tell us they don't like what we did. And creative confidence kind of goes down quickly unless we counter it with design school, you know, and ability to draw and watching YouTube videos and whatever. And you have to wonder if a lot of the new technology, especially the generative AI technology, which I'm sure we'll talk about, if that's actually restoring creative confidence for a lot of humans past that age. Now, if you have an idea and you want to make a quick animated short, you can actually make it.

4:13You know, it's like what an amazing bolstering of confidence that influences what we put out in the world. You said two things that seem to contradict. One is that creativity is this recycling machine, which does sound real, like we do just kind of riff on the things we've seen. And then you acknowledge that there's human ingenuity. How do those two balance each other? Better question, how do they exist in tandem? Right, right, right, right. You know, it's funny. When I advise product leaders on building new products, I always explain to them that a really successful breakthrough product is 95 % familiar and 5 % retraining.

4:50You know, if you try to do everything entirely new and new paradigms of user interfaces and new terminology and new icons, like the customer will just be totally discombobulated and they won't be able to use it. And so the trick is to make enough things that are familiar so that people engage and then find the one or two things that are completely different and differentiating and like do those in the entirely new way that trains the customer to do something differently, but also like distinguishes yourself in market. And in some weird way, I actually feel like creativity is somewhat similar because it's storytelling.

5:23At the end of the day, creativity that is effective is creativity that moves us. And we're not going to be influenced and emotionally engaged by creativity unless there's some degree of it that's familiar. So I actually happen to think that a lot of these things that we are picking up from the world, we are putting into the world the work that we put out there. And then the question is, what are you going to put in that's uniquely yours? And so that's actually the interesting debate right now about if I generate an image entirely just from my prompt, is it my creation? Well, if the prompt has a lot of ingenuity in it, sure.

6:01you know but if you do even more with that asset after you make it maybe even more so i think these are some of the questions we're debating these days right and we're going to get into the generative ai discussion pretty deep in a few minutes everybody waiting for that i promise you it's coming i also wonder whether you know the technology that's expanded our ability to do things has in some way removed a constraint that required us to have better ideas at the source of creative. So, I mean, this is kind of cliche, but let's talk about it. Like you look at some of the ads from like the sixties or the fifties, they're trying to sell a car.

6:38And it's like, I mean, the sort of the Don Draper days, right? The beautiful picture of the car with like a little, you know, short story about what type of person drives it. I mean, that it was an ad that looked and I mean, just sort of tugged at the heartstrings in a way that most advertisements don't today. Do you think we've kind of, I mean, you know, I think I know your answer because you're working at a creativity technology company. But yeah, make the case then for why all this tech has been good, given that we've sort of, it seems like we've overcomplicated a lot of our art today. Well, I think there's this question around the democratization of creativity.

7:16And as you know, I mean, any brand can create insatiable amounts of content now, whether it's blog posts or images or whatever else. And as a result, especially because of things like SEO and social media, every creator and every brand is flooding the zone as much as they possibly can. And so will that work or are we going to actually crave more meaning and more emotional engagement as a result of this flooding of the zone? You know, I happen to believe that two things are happening in the world of creativity. we're simultaneously lowering the floor so more people can get in and create stuff that ordinarily couldn't so you can make that animated short even though you don't have animation skills right i can make a a commercial without engaging a huge agency to do it we're lowering the floor that's exciting that's democratizing but we're also raising the ceiling so for those who are the don draper types and those who actually do have the skills they're not saying stagnant either Like they're able to cover far more surface area of possibility to find better solutions.

8:29And when you ask an amazing creative director or an amazing designer how they do their best work, it's always about time. Like they need cycles of discovery. And they always say, like, if I just, I mean, yeah, if I'm given more budget and more time, I will come up with more ideas and I will whittle them down to a few better options that will ultimately yield a better solution. It's just a function of time. And so in that instance, like this AI technology, as well as other efficiencies in the system, you know, should yield more iconic, more emotionally engaging spots. Yeah, I do think, you know, going back to your point about generative AI, putting technology in the hands of so many other people.

9:12I mean, the threads that I've seen on Twitter of people like dreaming up different images, whether it's like the president's looking like 70 rock stars or, you know, any matter of, you know, I don't know, just like cosmic imagery that goes on there. It's incredible because it's kind of interesting. You would never assign a graphic designer to make this. It would be a complete waste of time. And it does seem like this generative AI wave is really showing us something that we wouldn't see otherwise. I was talking to just like a little quick example. I was talking to a friend of mine who is, um, works like a big fashion label and his job is to come up with ideas for shoots, like, you know, for the, for the, for the campaigns and what have you.

9:56And he said that he had this idea one morning of a, uh, of like a, of a racetrack with like, um, racing cars. And, um, and for some reason, all of the drivers of these really antique old, beautiful cars were Japanese and they were wearing these scarves and he like, and they He had leather jackets and he had this like whole vivid image in his head. And so he went to one of these generative AI tools and just made an entire mood board of exactly what is in his head from like this, you know, morning dream sequence he had. And, you know, that became like a mood board that then they at work like looked at and ultimately refined what became the photo shoot.

10:37And it's like, you know, what an amazing toolkit that allows him to hopefully have an even, you know, better executed vision. So that's the kind of thing that I think will actually elevate the types of content that's made. Yeah, it's so interesting. Like when Dolly came out, I looked at it and said, you know what this really is? It's a communication technology. Is that something you'd agree with? Well, I do think that visual expression and storytelling is a very effective way of sending a message. And, you know, what are memes? Like memes are compression vehicles for knowledge and communication.

11:12It's one way for me to tell you something without having to explain it to you. Right. And, uh, and this technology, any, I mean, memes are creative, right? They're, they're images, they're visual. So yeah, of course, you know, I think these are, these are methods of communication. They just haven't been available. It was like a language that only certain people knew in society, like 1 % or less actually knew how to draw and graphically create something. And, you know, what happens when that becomes everyone to some extent? Yeah, definitely. I mean, for me, like it's, it's definitely become like the main way that I will draw the story art for my newsletter.

11:51And I was at a scale where it was just like, I'm going to be the only, if someone's making the art, it was me. Like it was going to be like a photo chop. Sorry to use the expression, but you know, like you guys must like that. It's people using this stuff in creative ways or, or, you know, pulling in a, a free stock image. And I think it's just enhanced the ability of people to get what I'm trying to say in a way that wasn't happening before and just made the stories more pleasant to read. It's pretty cool. I think it's true, Alex, but you know, we still need ideas though. It's like, it's interesting, you know, you can only think you're only as creative as like the prompt you can come up with also.

12:27Right. Totally. And, uh, you know, sometimes when you're thinking of a sunset, like, yeah, you can prompt a sunset, a sunset. But if you explore sunsets captured by famous photographers on a stock website, you're actually probably going to find a better image than you can even generate yourself. So there's, you know, there's like that, that question of, you know, can we, you know, it's still an imagination constraint equation. Right. And my mother-in-law is a graphic designer and I showed her Dolly when it first came out last year. And she was instantly like, oh, I'm toast. And I was like, no, no, it's not what's going to happen.

13:01Like actually these professional skills are even more in demand or that you can now communicate with people who want an image better and tailor it to, you know, sort of fit with their imagination. But now Genitive AI is also making its way into your products. So So I'd love to hear a little bit more about the way that Firefly has been baked into, and I guess it's Creative Assist being baked into Adobe products, and how designers are starting to use generative fill and other generative tools to make their work better. Well, we did start this journey by putting some of these capabilities out on a standalone website, more like a playground for people to play with.

13:47and what we immediately found was that people were excited, but they wanted to do it in their workflows. It was like, okay, this is cool. Novelty always precedes utility a little bit with new technology and it's a good way to test. But wow, if I could just do this when I'm doing this in Photoshop or if I could do this and turn it into a vector, but then edit the endpoints in Illustrator. Those are the types of things that we were hearing from almost all of our customers that were playing with this technology. And so then it became actually more of a interface and design innovation exercise. How do we bring some of these capabilities into the workflows that people use?

14:27And it's funny, like we have this product called Illustrator on iPad, which is actually not a hugely popular product, but it does, you know, it's Illustrator on iPad. And there was an innovation that the team came up with on that product years ago, because we wanted to eliminate a lot of the Chrome, like the buttons and stuff, because illustrators don't like that in their view. And so we came up with this kind of progressively disclosed bar, this action bar, where when you're touching, it gives you the options that are relevant for that moment, wherever you are in the workflow. And we leveraged that design pattern.

14:59And we took it into Photoshop to expose the same idea that wherever you are, we can give you the generative actions that are relevant for where you are in your workflow. So, and that, you know, that one innovation like changed everything for us in Photoshop. Yeah. So what does that look like in Photoshop? So in Photoshop now, when you are touching an image in your editing process and you, you know, and it detects that you could actually expand this image, it just gives you the option of generative expand. You know, if you, if you have a selection of something, you put a mask or, you know, You circle something in Photoshop, you then have the option of generative fill.

15:39And you can just prompt, like, I want to put in balloons. And then boom, you get balloons, you get many options of balloons. We just launched this vector, generated vector capability in Illustrator. And it's pretty amazing. I mean, you need an icon or you need a pattern or anything. And anywhere you are in your workflow, you just prompt it and get it. And it's, you know, it's, it just spans the possibilities infinitely to some degree. That's pretty cool. What are you hearing from designers? Like, is this something that's like immediately getting high adoption or is this, do you think it's going to take a little bit of time?

16:13Yeah. So the adoption is pretty immediate and, and there's pretty, well, let me put it this way. The adoption, you know, within a few weeks of us launching these capabilities and each of these products is higher than probably any other feature we've ever brought into these products. in maybe 10 years. So certainly as long as I've been involved with anything at Adobe. So I think the customers have spoken in that regard. But we, you know, we, I think it's still a question of like, where else do we integrate it and, and how do people use it? And, you know, but I've never met a creative professional that wants to take two hours to do something that they could achieve in two minutes or two seconds.

16:55No one wants to do things the long way. The question what are they going to do with a lot of their newfound time? Exactly. I mean, it's kind of basic, but the resizing alone is amazing. I feel like designers must spend half their day taking that one image that was square, making it horizontal, making it vertical, making it poster size. And if you can use generative AI, I mean, you talked about expand to expand images, to expand them, to contract them, to produce all those assets in a snap of a finger, you're all of a sudden left with much happier designers would probably make the whole organization happier.

17:31Totally, because they don't want to be doing this mundane, repetitive work. And when you give them that time back, you know what they're going to do instead? They're going to like run more A-B tests of different images as opposed to resizing the same images. They're going to start to explore other possibilities and propose maybe better solutions. So I think that's the other thing that I've been involved with a lot of these conversations with companies that are asking this question of, oh, is this going to give me more efficiency in the creative team? And what does this mean in terms of do I have to hire less creatives?

18:01And I always explain, first of all, engineers have gotten more productive for the last few decades. And guess what happens? People keep hiring more engineers. Why? Because they get more ingenuity per person and they want more people because they want to do more things. And unless you're a penny pitching in a private equity-owned company, you are probably a company that actually wants to build more products, more features, more marketing campaigns across more platforms and more regions with more testing. And so when you get capacity back and actually get more out of each person on average, actually you end up wanting more people because you end up wanting to do more things to grow your business more.

18:43It's sort of like maybe just a natural human tendency of capitalism, if not just humanity overall. Yeah. So let's talk a little bit about who owns these images. So it's kind of interesting, right? I'm just thinking through the path here. So someone makes an image, that image is used to train a model. Someone's making another image that used generative fill to enhance it. Do they own the image? Does the person who previously created that original image that was used to train the model own like one millionth of their image what what is your what's how do you think through this stuff at adobe well a few things first of all i'm not a lawyer as you know alex yes i spend a lot of time with them these days but i'm not a lawyer myself uh and um and and also the the the rules and the laws around some of this stuff are still in some areas either not fully evolved or tbd and And it's also like different regions and countries are exploring this in different ways when we get into all the evolution of copyright law and everything else.

19:52Currently, you know, there's also a difference between what we are doing and why we're doing it versus what we're legally required to do. You know, I think we've been a little bit more creative, protective, you know, because A, this is our customer base. And B, we believe actually that this is the way the world's going to go. But some companies are taking a different approach. and would disagree with what we're doing. When you make something through generative AI, there's a lot of legal thought that when you edit it in any way, it is yours at that point. And so that could be as simple as, if you're in something and you're expanding an image that you took, it is your image because you actually generated a lot of it.

20:38Now the question is, is it 50 % of the pixels? Is it 99 % of the pixels? And I think you pose this good question of, well, at what point is it an immaterial level of input or editing that you're making, you know, to this to be able to claim that you made this image? And I think that's, you know, that's a great that's one of those like open questions right now. If you, you know, there's some folks that say that if you entirely generate an image just with a prompt and one of these tools that was trained on other people's content, it's not yours unless you edit it further. And that's, you know, again, though, I don't have like a legal position on that per se, but those are like the debates that are happening right now.

21:16Yeah. So you're writing on Substack. Did you know there's a Substack feature that allows you to toggle on and off whether models can train on your work for generative AI purposes? Have you seen that? Yeah, I just actually noticed that yesterday. Funny enough. Yeah. Okay. So what's your choice on that? It is wild. And it's like, wow, this has become all of a sudden, like, I mean, if you told me two years ago that I would like have an option in the newsletter tool that I'm using to let AI models train on the work, I'd be like, what are you even talking about? But we're here. Well, I was like, first of all, I had some funny ideas, you know, before that feature was discovered, you know, I was thinking, you know, what would happen if I wrote in the end of my newsletter?

22:02and by the way, everything I just wrote is fake and this is incorrect information. And then I put like a wink, you know, to the, you know, to the AI models that are training off of this, like with that, then tell the model to actually ignore what I just wrote because it's fake. And, you know, is that like a way of me sort of protecting my content from being trained? Like there's all these sort of mental games now I play. And also, you know, do you, is this how paywalls are going to become super popular because there are certain things you write where you're just going to want them like not on the not on not trained and and you want people to have to log in and be able to read it as humans so there's there's all kinds of questions there so what did you choose um i um i chose to allow training um but i also have a section at the end where i have uh like i share some more provocative thoughts and like things that i'm working on And I put a very low paywall in front of that, that I do that I give the charity.

23:06But I figured that was one way where I could just like make sure that that stuff, you know, can't be accessed by a bot. Yeah. Wait, do you think the bots can't access underneath the paywall? My understanding is that they're not able to scrape it. I guess someone who's like paying manually feeds it to a bot, but I thought it was a safer approach. Yeah. So I also chose to let it crawl. And at first I was like, no, no, no crawling. And then I was like, wait a second. If someone is talking to a Bing or a ChatGPT and they want to learn about technology and the bot is pushing big technology, I'm like, yeah, of course.

23:46It'd be the best advocate I could possibly find. Especially if attribution becomes commonplace and you can see where something came from and you can follow up and learn more about it. Exactly. All right, let's take a quick break. Come back after this to talk about some of your writings and implications and a few other questions about the generative AI business. We'll be back right after this. We're here with Scott Belsky, Chief Strategy Officer and Executive Vice President of Design and Emerging Products at Adobe. See you right after the break. Did you know your credit card points and miles can lose value to inflation?

24:19Credit card companies often reduce the redemption value of your points and miles. Now, imagine a credit card with rewards that can grow in value. With the Gemini credit card, you can earn Bitcoin or one of over 50 other cryptos instantly with no annual fee. Every swipe at the store or gas pump earns you instant rewards deposited straight to your account. Plus, sign up now for a$200 Bitcoin bonus to kickstart your rewards. Visit Gemini.com slash card today. Check out the link in the description for more information on rates. Again, if you're looking to invest in Bitcoin but don't know where to start, the Gemini credit card makes it easy.

24:57The Gemini credit card is issued by WebBank. In order to qualify for the$200 crypto intro bonus, you must spend$3 ,000 in your first 90 days. Some exclusions apply to instant rewards in which rewards are deposited when the transaction posts. This content is not investment advice and trading crypto involves risk. The Gemini credit card cannot be used to make gambling related purchases. What the hell is going on right now? And why is it happening like this? At Wired, we're obsessed with getting to the bottom of those questions on a daily basis. And maybe you are too. I'm Katie Drummond, the Global Editorial Director of Wired.

25:34And I'm hosting our new podcast series, The Big Interview. Each week, I'll sit down with some of the most interesting, provocative, and influential people who are shaping our right now. Big interview conversations are fun. I want a shark that... That eats the internet. That turns it all off. Unfiltered and unafraid. So in a lot of ways, I try to be an antidote to the unimaginable faucet of reactionary content that you see online, to the best of my ability. Every week we're going to offer you the ultimate luxury of our times, meaning and context. True or false, you, Brian Johnson, the man sitting across from me, one day, at some point, as of yet undefined in the future, you will die.

26:18False. Tell me more. Listen to The Big Interview right now in the same place you find Wired's Uncanny Valley podcast. Subscribe or follow wherever you get your podcasts.