Everyone's an Artist: What AI Generated Art Means for the Creative Process (VIDEO)

Austin Mace, our chief innovation officer, recently sat down with Shep Bryan, our strategic growth lead, to chat about AI art generation tools. Spoiler alert: we love them.

‍Connect with Austin and Shep on Linkedin for more metaverse content.

The transcript below has been edited and condensed for clarity.

--

Austin Mace: Hey, Shep. How's it going?

Shep Bryan: Hey, Austin. It's going well. Excited to be here today talking about AI generated art with you.

Austin: Awesome. Thanks for coming. So, let's jump right into it with some quick history. Over the past 10 years, we've really seen a lot of evolution in what's happening with GPUs or graphics processing units. In computer terms, a lot of things are done on the CPU, which is your main processor. But over the past 10 years, we've really found ways to leverage the GPU to do things beyond just generating video images for games. From simulating traffic patterns, to weather patterns, to speech patterns, now we've found a way to apply it to images. So there's been a really cool combination of pairing that technology with cloud computing and text-based prompts. So, Shep, can you give a quick rundown of what you've been seeing in the field as of late.

Shep: Yeah so for starters, to answer this question of what are AI art generators, basically, we've got a machine learning algorithm that's being trained on huge amounts of data sets. A lot of these are publicly available, so with a tool like Dall-E or Midjourney, you are pulling from what basically amounts to a Google search for your prompt combined and collaged together to get a different set of creative outputs that's fully unique, fully generated through this artificial intelligence, machine learning tool that ultimately just makes art.

Austin: That's awesome. I know you were mentioning some examples, including Midjourney and Dall-E, so what's exciting you most about Midjourney?

Shep: So Midjourney in particular is this beautiful combination of most advanced yet accessible. You've heard me say that phrase all the time, Austin. With a tool like Midjourney, all you need access to is Discord because they've built this as a Discord bot. You go into a Discord server, you type "/imagine", and then you type your prompt. And that's all you need to do to actually output these beautiful, beautiful visuals. Sometimes it's a little bit more like nightmare fuel than beautiful visuals, but that's what the human driver of the AI tool is for. So Midjourney is pretty exciting. It makes it really easy for anybody to access this amazing new technology.

Austin: That's awesome. So, I want to play a little game today called Guess the Prompt. And I'll even let you have a little wiggle room. So guess the prompt or guess the pun or play on words. But we do have a theme today, which is Titans of Industry. So we've got four lined up for you. So let's roll into the first one.

Shep: Alright. So my guest here is going to be Jeff Bezos Flying High.

Austin: You know, that's close. And this will be a pattern that we'll see over the next ones. This is Jeff Bee-zos. So the prompt was Jeff Bezos as a bee. So alright. Let's move on to the next titan of industry that we've got lined up here.

Shep: Let's see here. Well, we've got an apple, right? Let me put on my hat and think like AI here. Maybe this is Steve Jobs as an Apple.

Austin: You know, that's very close. This is Tim Cook. Tim Cook as an Apple. So we got we've got Tim Apple, as he was famously called a couple of years ago. Let's move on to the next one.

Shep: Elon Musk as a Twitter bird.

Austin: So close. Actually, not really. This is Elon Muskrat. So he's got the muskrat hair. Yeah, there's a little bit of a theme here. Alright. Let's see if you can get the next one.

Shep: Oh, boy. This is ... I can't remember his name. This is the guy that runs Virgin, and we're riding some kind of first-class buffalo in the sky to reach our destiny.

Austin: I'll give you the first name: Richard

Shep: Richard Branson's first-class bison to space.

Austin: It's Richard Bison. So close. So close. Alright. Well, let's end strong on this one. Who is this social media magnet?

Shep: This is Zuckerberg Cucumber.

Austin: Mark Zucchiniberg. Come on, man! Well, you know, to be honest, you didn't do too great, but that's OK. I think it was fun exercise anyway.

Shep: So maybe you could say, Austin, that my intelligence is artificial.

Austin: Haha. Nice. No, I wouldn't say that. So, Shep, what do you think AI art means for the future of creativity? What are you excited about?

Shep: I am absolutely thrilled with the horsepower that this puts behind the early stages of creativity. When you sit down with a blank piece of paper, sometimes it can be hard to fill it out. But the magic really starts once you've got those sparks to work from. And the fact that with two words you can come up with some really, really novel visual concepts that can really spark a brainstorm pretty quickly is great.

Austin: Yeah, absolutely. And we actually have some examples. We started to use Midjourney here at our studio to brainstorm some concepts and have some fun. I think some of these examples are yours as well. So can you talk through what's been your inspiration for some of these?

Shep: Yeah. The one behind you is God's Pizza Oven in the Sky. I was cooking some pizzas last night with friends and sometimes somebody says something and you just want to see it as a Midjourney prompt. The one in the center is, I think, just Iridescent Skyscrapers. And one of the neat things with the process of getting creative is once you put your initial prompt in, you can re-roll it in Midjourney. So you don't need to come up with a crazy novel prompt every time. You can iterate. If you're imagining a decision tree from your first prompt, it gives you four sets of images like you're seeing in the center of the screen here, and you can pick one of those and kind of go down the rabbit hole. There are a lot of powerful words in the English language that have visual information associated with them. Things like "iridescent" and "skyscrapers" and celestial ethereal words like that have a very strong visual language, and in many cases the AI is able to pull some really compelling creations out of that terminology to make something like what you're seeing here.

Austin: Yeah, 100% And the two here - the Eiffel Tower and the bean here - I can't say what pitches they were for, but we actually put these in front of clients and had a good 20 to 30-minute conversation around ideas that we were presenting. And that was a pretty huge Eureka moment that Midjourney could be used in the creative process to help convey ideas. And so yeah, sometimes you get some weirdness out of these, but also, if a picture is worth a thousand words, what's an AI generated picture worth when it comes to conveying an idea? So, to kind of wrap things up, do you have any parting thoughts? Where do you think this is going to go? Do we think are we going to have Netflix shows writing themselves? Are we going to have more accurate prompts coming out? What do you think?

Shep: I'm fascinated with how this is empowering the creative industries, not only in visual art. You've also got the same thing kind of happening with music; AI generated melodies and song structures. I think if we're thinking about how this can be used as a tool to improve creative processes, the simple answer is figuring out how you can create an actual strategy for empowering the creatives on your team. To use this as an ideation and an inspiration tool, I think we're very quickly going to be seeing how these tools emerge as a really functional part of the process. Midjourney already has a corporate license that you can sign up for if you're an organization to use this as part of your institutional processes. Austin, I don't know if you have any thoughts, but there are definitely a lot of exciting things happening here.

Austin: To kind of wrap it up, what I found as a creative and someone who works with technology every day, is that it's definitely not something I've perceived as threatening. I think there's still going to be a massive need for folks in the creative industry, but it's really nice to have another tool in the tool belt. I, for example, I can't draw very well. So being able to, with my words, write a prompt that generates an image to then help convey an idea has just been such an amazing feeling. And so I'm really excited to see where this goes. That's all we've got time for now. So, thanks for dropping in, Shep. I really enjoyed it and thank you so much.

Shep: Cheers, Austin.

Austin: Cheers.