AI and Ethics in Architecture

A Welcome to Practice Disrupted, a podcast where we find ways to create new solutions to current challenges while elevating the value of architects. I'm your host, Evelyn Lee, an architect who spans tech as an angel investor, startup advisor, and founder of Practice of Architecture. Whether you're a seasoned architect or just starting in the field, this podcast is your gateway to think differently about the role architects play within our global community. Hello, disruptors. Welcome to this episode of Practice Disruptive. Today, we're diving into the world of AI for image generation and exploring the ethical implications for the architecture profession. Joining us is Damon Leverett, AIA, NOMA, an architect, artist and educator with a unique perspective on this topic. Damon is an adjunct lecturer at the University of Arizona, teaching courses like architecture design, installation design, and computing in the arts.

A He also co leads the university's AI Industry Working Group. With over 35 years of industry experience and former roles with the AIA National component, NAAB, and leading firms like SmithGroup and Page, UIP, Damon brings a wealth of knowledge. He recently established his own interdisciplinary practice, Lever Arts, focusing on design, art and technology integration. Damon is also the 20 24 past president of AIA Arizona, further underscoring his leadership in the profession. In fact, his seminar on AI applications and architecture was recently accepted for the AI 8 24 National Conference. As AI image generation tools like Midjourney, DALI, and Stable Diffusion gain popularity, they present new possibilities but also raise important ethical questions for architects. How should we approach leveraging these AI capabilities? What are the risks and implications that we really need to carefully consider? Damon is at the forefront exploring these issues. We'll tap into his expertise to better understand the responsible use of AI image generation, the impact on architectural practice, and much more.

A This is sure to be an insightful and thought provoking discussion. So let's dive in. Hi, Damon, welcome to the show. I'm so glad to have you here. I can't actually trace back the origins to our friendship since we had a relatively long history working together within the AIA. But I've been interested in seeing how your path has evolved and changed even, you know, within the last 5 to 7 years. And we like to start off practice disrupted with a bit of an icebreaker. So we opened with your formal bio, but is there anything that you can share with our listeners that may either help us set up this conversation or just might be an interesting tidbit that most people don't know about you?

B Well, I think one of the things that most people don't know is my love for athletics. And I, was an avid basketball player in high school as evidenced probably by my height. But I the interesting thing is I had to give up organized sports to go to architecture school. Mhmm. And, you know, it's just something that you couldn't really do both of those at the same time. So I I don't have any regrets at all about that, and I'm actually continue to be involved in all kinds of athletics, either, bicycling or swimming or or a little bit of running here and there. So I think I just want to say that being engaged in athletics and competition has really helped me in my career. It's taught me leadership and, teamwork, and I think that's been really important.

A Yeah. No. Thanks for that, Shah. I don't think you and I have ever talked about athletics from that point of view. Now that I quickly run through all of our past conversations in my head. So I appreciate that. Why don't you tell us what you have been up to most recently, especially what you're most passionate about with with AI too?

B Well, thanks for asking, Evelyn. I I think the one thing is important to me is that I've experienced another disruptive technology in my early part of my career, which was which was the shift from manual drafting to CAD drafting. So I I saw the anxiety that happened both in the firm I worked with as well as within the general industry than in the city I worked in. So I wanted to get involved in this right away because I feel with my experience in in transition technology transition that I could dedicate myself to really helping others get to to know this technology, artificial intelligence, and to provide some training, which I've I've been doing. So that that's really what has been important to me is really understanding both of those dynamics, the past and the present. Sometimes I tell people that if you are talking about AI and all the difficulties or all the enjoyment, you could just replace the word AI with CAD and then roll the clock back 30 years, and then you would see that it's basically all the same questions, all the same inquiry. So I think we're real really trying to, get ahead of the game here.

A Right. I think you have, and particularly, have been focused a lot on on imagery. At least that's what gets posted a lot out to social media. Are there any other aspects that you've been kind of researching in your own education on the topic?

B Well, the interesting thing about artificial intelligence in the last year or so is that there are several people like myself who are active in education and lecturing and so forth. But I think all of us find that we cannot cover every aspect of AI. Sometimes audiences and groups will actually feel they can ask us any question about any anything. But I've decided early on that I would just focus on really just 2 things. And one is, imagery and the other is ethics and risk. So I'm doing a conference lecture with the, AI construction documents committee. I just did one with SMPS. And, basically, those presentations are focused more on practice and how you would integrate AI functionalities into your everyday function as an architect office.

B And so those really are the 2 things I I focus on a lot. I I enjoy so many others. I enjoy chatbots. So I've actually written a story using chatbot, a 10,000 word story, and that was a lot of fun. And, I experiment a lot with everything I can get my hands on, but I I have to stay focused because, I just can't do everything.

A I was gonna say you still picked like 2 very large topics, I'm sure you can do an entire semester long course right on just AI and imagery relative to architectural design, and then AI relative to ethics and risk, there's there's definitely no shortage of information, even within those 2 subtopics. So let's go into I thought, you know, you've been traveling all over the US or even, you know, virtually, globally through these online presentations. And I thought it would be helpful to go through some of the most frequently asked questions that you're getting in these webinars, because I feel like, obviously, there's themes there. Obviously, I feel like they they're kind of the most topical when it comes to architects deciding whether or not they even want to experiment with this new technology. So why don't we just jump right in and start with the big one? What's your take on will AI replace architects?

B That is probably the most common question I get over the past year from different lectures and conversations. You have to think about it in a historical point of view. Just remember what happened when the automobile replaced horses and buggies. There was a lot of anxiety, but in the long run, there's a couple million people in the automotive industry, and we know how valuable increased capabilities of transportation is and what it has done for society, transforming society and eventually becoming airplanes and so forth. So the benefit of that technology disruption has been enormous for humanity, and it's been worth it. And in fact, some of us feel that every situation does not necessarily mean we will lose net jobs. Some will be neutral. And in many cases, like the automotive industry, it will actually increase the number of jobs.

B So it's it's actually I think when people ask me that they think of it kind of as if there were a robot came into your office and started drawing everything. And it's just really not that right now. And maybe in the future, there will be some advanced form of of AI that will do all kinds of things. But mostly, just like robotics, AI is doing the more repetitive, mundane tasks, so it leaves the professionals to be able to be more creative, to use the time in different ways to advance their their objectives rather than be bogged down on the nuts and bolts.

A It's interesting to hear you say that because I also think, you know, with your imagery to generation, you are using AI From a creative standpoint. Let's talk about the tools in general, what tools do you tend to gravitate to? I feel like some architects go with DALL E. Some architects are playing just in in Midjourney. What tools do you tend to use and and why?

B Great question. I I honestly have tried to use as many as I've can I can? So I've used DALL E. I've used Stable Diffusion. Chat g p t has its own version within chatgpt4, which is actually a connection to Stable Diffusion. So there are so many, Leonardo, Google has 1. Microsoft has 1 also connected to its browser. So there are dozens of these out there, but I've tried enough of them to know that midjourney just has a sense of quality to it in terms of its rendering, its response to the prompts that I add to it. I just can't switch to something else.

B It's just really superior. Just in I know that a lot of people feel differently. They they become accustomed to a different system, and they feel pretty strongly about it. I I know a few people who are really focused on unstable diffusion, but, I really feel midjourney is is the superior tool at this time.

A Yeah. I'm still working out. I I mean, honestly, even as somebody who is tech forward as me, and I haven't tried stable diffusion, The fact that you have to access midjourney through discord and kind of learn a little bit about servers within discord itself was a bit of a roadblocker for me even on getting started with Midjourney.

B Oh, I've I've got good news.

A Oh, okay. Let's hear it.

B So currently, I'm an alpha tester for MidJourney's new web interface. And, obviously, it's not proprietary for me to talk about it because, it's basically out there with probably several 100 people doing this. But the important thing is is that Midjourney has not announced when it will actually go live with the, product. But I can tell you right now, there's no more Discord. It's easy to use. It's just a web interface, and it's far superior in terms of its flexibility and how you can blend and and add images and change text and do prompting in all kinds of strange ways. So I I am really super hyped about this new product, and, I've actually talked about it in in my writings on the on the web. And I will be updating people as as soon as I hear more.

B You know, the thing is is people have complained about that, and the company heard everybody. And they from what I can see, they put a lot of effort into making this right, and I'm I'm really excited about it.

A No. I mean, that's really exciting. I think anything that you can do to remove blockers to increase adoption of these technologies, I think will be will be helpful going

B forward. If I could also also add that if you're like me and you want to do more interactive learning scenarios with, adult learners, it's very just think how difficult it is to to teach Midjourney when you have to show them how, you know, how to get a Discord account. You just can't do it on the fly. You would have to prepare everyone a week ahead of time. And, with this new interface, I'm I'm excited because it'll make it easier for me to teach other people to use it.

A So let's talk a little bit about prompting. You know, there's there's whole certificate courses that you can get out there now from universities on on prompting, how you get people, if they're on the platform, if they're in Discord or, you know, this new version of MidJourney, how do you tell people to get started in exploring this medium a little bit more?

B So when I'm working in a lecture environment, I have several slides on this. So what I do firstly is explain the process of how AI creates an image in general, but then one one part of that section is on prompting. And I codify it this way for people who are starting. There's 2 ways to approach prompting, especially at the beginning. One way is called conversation. In other words, to describe a picture you want to create, you would just write a paragraph as if you were telling amazing story to your best friend. So I've seen people do I've seen people do this, and it sometimes could be a 100 words. But there's nothing wrong with it.

B And what it is is a great learning experience because you could begin to see how AI responds to the different things you say. The other approach that I I basically teach is called the ladder approach. And, essentially, this is what I use even today when I build imagery. And that is to start with as few words as possible, understanding the basics such as starting with general terms and moving to specific. I always tell people that think of it as that you're in the universe, we're in the galaxy, we're in the solar system, we're on earth, we're in the state we live in. It's if you think like that, it's actually becomes easier to see how the logic of the computer can follow that. So you wouldn't start with the color of the sky when you really wanna talk about a dog house. So you would start with the dog house first and then start to develop the description of the dog house and so forth.

B So that's mainly how I I do it. I'm really interested in in how things respond, and, you know, so far so good. I should also mention, you asked this question too about, a little bit about how the words you have to put in words to make this happen. That's a concept sometimes we refer to as alignment. And alignment is the intentions and their outcomes. It's the it's the dialogue between those two things. And what happens often is, many users become frustrated because they cannot get exactly the words to do exactly what they want. So AI is this it's a limit to what AI specifically, this version of AI, which is narrow AI.

B It cannot really do a great job of making that correlation perfect. But what I tell people is that you create so many new things, so many incredible novel ideas that it becomes a different way of thinking about design and creating. And you have to leave yourself a lot of room, have an open mind, and be able to adjust your process to respond to these, sort of risk the outcomes that AI will provide to us. So it's just a different way of of looking at it.

A Yeah. No. I get that. So you're you're touching a little bit on on the creative process. So so this is an interesting thing that I've actually heard a few females say, it'd be interesting if you heard the same notion. You know, what if people feel like AI is cheating, like, I'm cheating the creative process by by using AI, or I'm somehow circumventing natural human creation by leveraging a tool like AI.

B Yeah. I was in a lecture where I shared the stage with, a professor from a major university, and he said that he would not allow his students to use Midjourney because in architecture because he felt that it was too easy. So we actually have a contrast between it's not like we really disagree, but I actually use an assignment. I teach computing in the arts at University of Arizona, and I actually have two assignments in the class that I teach people how to do. My I teach students how to do imagery with with AI. So I I don't have a problem on them doing it. So a lot of times, this is how I answer that question about that concern. And and that is is that, the example I give is is this is fairly well known.

B That if you think of all the medical research papers that are written in the world every day, it's a sub substantial number over a 1,000. We'll just use the word 1,000 because I know it's far more than that. 1,000 papers per day that is, you know, 365,000 papers on on medicine. And if you think about it, there is no human you know, if you take 5 years, there's no human who could read all of that and synthesize it and perhaps in the search of finding, say, a cure for something. And and so when you look at the downsides that some folks are articulating, we have to be optimistic because there's so much that AI can do for us that is possibly not available to us manually. So it it's an amazing thing. If I could discover through AI a new solar panel that was 95 percent efficient, would AI be worth it? And when I ask questions like that, people say yes. So it makes the conversation a little more balanced in how we could think about the things that are disruptive or maybe don't fit my process versus the potential for all this amazing discovery that can happen with AI.

A I'd be interested. What was the professor's response to to that response?

B Well, I I don't wanna pick on him, but I think that, you gotta look at the context. Let me put it like this. There's probably about 20 of us going around the country talking about AI and design and architecture. It's probably less than 20, really. I know I know the books of the major people in the universes they they're from. But, essentially, just the 20 of us could never reach, you know, the whole audience that's that's necessary to make that happen. But I have only been engaged with my cohorts, if you will, in 2 engagements where we are actually presenting together. The key is who goes first.

B I hate to say it like that. So, you know, I went second. So it really isn't fair for me to

A Right.

B You know, I I wouldn't I wouldn't actually critique what the person before me said because it just it's sort of rude. But, you know, and maybe in some ways, when we're actually talking about a topic that requires more civility than probably anything we've ever known, you know, except dealing except dealing with our family or something like the only thing I compare it to. You know? So we sat down and talked about that was, like, the right forum. And I think he he saw my presentation. He saw how beneficial could be in the design process, and I think maybe perhaps he was maybe considering something different. And I hope he does. Because if you teach AI imagery in college to architecture students, you could put the guardrails that you need around it to make it work. So one of our professors at University of Arizona has asked his students to use it as inspiration.

B So that's really the key, and I may be getting into the next question a little early. But you can use it just to inspire yourself, or you could use it as the core design process and actually one little element in the path between a notion all the way to construction administration. And, you know, there's so many parts between those 2, but we have to find out find out where AI fits into that. So it it's really is a chance for us to think about AI and how it fits into the to the process.

A What is though, like the intersection between AI and and the real world? You know, if I were doing a studio, but I actually wanted a site for something to exist on, Is there the ability to translate my my renderings to to a a site map and and geotag it and and things like that?

B Well, sometimes what happens when I do lectures and engagement with design professionals, they ask me lots of questions. I would say about 25% of my answers are, that sounds like a great idea. Why don't you try it? Or I'll say, I'll after this lecture, I'll go home and I'll try it. Because that's the nature of this tool. It's not really that limited to just typing in your ideas. All you can do is try it and see. Now in particular, someone asked me if I could could do that, place a building on a specific site in AI as opposed to, say, putting it in a Photoshop version of and then meshing my AI with Photoshop. I think what I did was I I went home afterwards and I tried it, and I found out that it really can't do that.

B So I found a very quaint street in Paris and found the address and got the Google Street View and everything, and then I described the location with the longitude and latitude and the street name and the city. You know, how much more specific could I get? And, the thing is is that it still can't do that. So what it did was create an average of buildings around there that are basically formed around all the buildings in the neighborhood. So they're fictional buildings that are created in AI, but they're not the actual buildings on that street. And so I wrote about that. To answer that question, I wrote an article about what really happened. And I I want to do that and, you know, I'm really considering soon to post to take request where people have questions like that, and I'll take requests, and I'll try it. And I'll just hammer hammer it as hard as I can to try to do these things and and then report back for people what happened.

A Yeah. To your own credit, that just goes to, like, how much more experimenting and figuring things out there is left to do, right relative to, to how we use this technology and how we how we fit it in. What's the next iteration, You're using it very creatively to create these wonderful images that, you know, if anyone follows you on LinkedIn or Instagram, they'll see. Would you ever put that imagery in front of of a client knowing that you used AI to create it?

B First of all, AI is being used in architecture firms right now. So I've talked to a few firms, but it's not widely known how they're how they're using it. Because, unless they publish something, then some people ask me why that is. And that's because of one of 2 things. 1, they wanna keep it proprietary. They don't want either the competition to know how and what they're doing. The other issue is they haven't built anything related to to AI. And the other the other one I sometimes think of is they're embarrassed to say they're using it.

B So this one of those three things. Alright. It's kind of not embarrassed. It's really not ready to admit that they're doing something like that because they're concerned about the reaction from the public. They don't know what the reaction is going to be, and they and they can't really be very loud about it until they know it becomes more of a mainstream topic. I do know of a architect who told me a story where they took a bunch of images and put it on the table to do it in what we used to call image analysis. So image analysis is not done by every architect and some architects are opposed to doing that. It's basically just giving your client a feel for what the building could look like.

B And it's really no different than if you were doing a museum for a client and he said, I want a really great museum. When I would say, let's get on a plane and go see the best museums, and we would just look at them. And sometimes we use that inspiration, and the client will say, hey. I really have this not that he likes it. It's not a like, dislike, but he says, I have an emotional response to this particular building. And it helps inform the architecture team about the direction they're headed. So if you think about it, AI imagery fits into that role very well because you can create almost anything that you want in terms of architecture and use it in that scenario to help communicate. And so the architect that I heard about this is essentially didn't expect this to happen, but he laid out the 10 images to kind of show what they were thinking.

B And the client really didn't know how they did those images, I think. But essentially, he said, okay. Yeah. That one. So by doing something like that, the team had to go back and figure out how to do that. So that that falls into 2 categories as I I said before. 1 is to use it as inspiration, and the other is to use a brute force design. So brute force design is taking a image that you created in midjourney and then reverse engineering it in 3 d modeling to make a new building based on what you see.

B And I've had I've done this before myself, and it's not that easy. But, for example, there's some obvious things going on here. One is when you do an AI image, you can't see the back of the building. So it leaves you a lot of design leverage so that your team can actually design the entire building and accept that some parts of it are actually created in response to what they see as opposed to what they see. So that that's actually a process we're working. The other thing I think it falls into this is is something I get a lot when I I speak is that I hate to say it this way, but some people will say, oh, those are just pretty pictures. And that's not really architecture. So I don't I don't get heckled a lot, but every once in a while, I'll get that one.

B But the the main thing is is that there are some limits to where AI is today. Now I have to give a caveat that when I say there's some limits to x, y, or z, that's only today. That doesn't mean there's so much AI happening so fast that literally tomorrow, someone might figure out how to do x. So one of those x's are doing floor plans. And the key element to that is the difference between raster tools and vector tools. So, essentially, midjourney is an image. It's a raster tool. It's just pixels showing a depiction of a of a of a building.

B When you have a very powerful vector tool, it can take your prompts and actually create line drawings from that in an accurate way. Now there are a lot of tools who leverage this. Some of them are not new that can do planning. They can plan residential buildings or they can plan the whole neighborhood or they can plan the site. And I won't say that there doesn't exist a version of this that can really generate a plan from text, so we call it text to plan or something or text to architecture. But, also, maybe next year, someone will figure out how to do that. I just personally have not seen it yet. I I've tried to do some of these things, and I've tried to do floor plans in a system like Midjourney, which is, again, is a raster based, product, and it really cannot it cannot do it, very well.

B It's it's not really saying that the system has a shortcoming. It's just the the question we're asking is so complicated, and the requirements that we need in order to say make a floor plan work. You know, there are things such as there's a closet in the bedroom, then there's a door, and and and then there's a pathway. There's if you think about what we do as architects, there's a zillion of things like that that we now we can do unconsciously and with the assistance of codes and other regulations. But there's so many pieces of information involved in that that it's very difficult to input that into a raster based system.

A What about ethics, you know, relative to to AI? You've obviously talked about it being mostly applied to the creative the creative aspects of what we do. Is there any risk to using it really right now?

B Well, I have a lot of information, and I've actually done some serious research on this topic. And there's a lot of anxiety, about using AI. And, actually, to further the conversation, I'm actually about to give a presentation to a group of artists who are really concerned about that. So there's several things going on. One is is that all the products we create in Midjourney as imagery are not copyrightable. So that's the first issue. The US copyright office won't grant a copyright to, element that was not created by human was not created by humans. So that has been a small limitation there.

B The second thing is that it's very obscure as to how training models as we call it, the things that build the information and data that can create something new, how that model was created. So the average person walking down the street has no idea. I I there are tons of people who are are really good at machine learning. I I actually teach with a few of those folks who I have tremendous respect for the kind of science that they actually put into that that kind of programming and computers. But there there isn't really a a way for us to know exactly. So some, I had an artist come to me and say, well, that's just stealing from another artist. So I also go through an explanation that is it's the way I explain it is if I had a 1000000 pictures of cats, and I took one pixel out of each one and I use those 1,000,000 pixels to make a new picture of a cat that's totally different, is that really appropriation or not? And so most people when I say it that way, they would usually say no. However and this leads to this overall principle that one size of regulations and concerns does not fit all types of AI products.

B So literature, art, robotics, science, they all need different ways of looking at them, and this is a really good example here is when you think about AI in the sense of literature. So if I were creating a chatbot that and I asked a question about, say, the World Series, where does it get that information from? Now the claim that's going on right now in a court case says that they're stealing it from the look you know, the local or national newspaper in order to do that. You could think of it this way as a different from the one pixel scenario because no one can recognize a pixel that belongs to someone else, especially if it's 1 millionth of a pixel. So it's not an issue, I think, in my opinion. However, if I were writing a article about the World Series, and I said that, well, I bought one letter from an article from newspaper x. Is that is that okay? Well, most people would say, yeah. That's fine. You borrowed one letter.

B I mean, it doesn't give us any information. It's pretty obvious. Now if I said, what if I stole one word from another newspaper like one word? Is that okay? And the answer is most people will say, well, that's probably fine because you can find words anywhere. But when I say, what if you could take a sentence from newspaper x and put it in your article? Is that okay? And then most people will say, oh, you know, I think that you would have to cite that. So that's why the issue of literature, you know, and design imagery are in different places. And I just used that conversation to tell people, especially when they're looking at processes in their office, that they have to look at every process differently. So there's CDs, there's emails, there's letters to the client, there's field orders, there's change orders, there's contracts. There's all kinds of things like that.

B And each one of those elements needs its fresh look at how risky it is to use it. And I think that's the really the lesson I'm trying to get out of that.

A So no one size fits all. And, obviously, a deeper understanding of the tools that we're using to understand how how it's affecting our output and and where we wanna use it within our own design itself.

B Exactly. This is what I say about the optimistic viewpoint. If we get bogged down too much in the risk, if we tip the balance so deeply to one side of that equation. Some people will say, well, what's wrong with that? You know, we're safer. There's lots of ways to describe what safe means in this scenario. It's pretty obvious some of that. But, really, what it means, if if we do too much of that, we will stifle innovation. And that is really unfortunate that if we did that now I know at certain spheres, like, one of the things that I I have a foursquare, and I talk about things that are really risky and visible.

B So, obviously, like, medical information is one of those things. So we could say we have to have privacy. These are all things that are really important. Privacy, the ability not to harm people, not to discriminate against them, not to abuse the information. But at the same time, if we can discover a new drug that will cure x, then we have to balance those 2 together to make sure that we can advance humanity and not and and not stop innovation. So it's a very tricky thing to do. It absolutely is a very exciting time to be in this condition, really, because there's so much potential, but we have so much work to do in order to be civil to each other as we do it. So I think it's, it's a a very good, balance of, ideas.

A Well, let's talk about work or the use of the word work since you just since you just use that. I do believe there's a lot of work to do. And when a lot of people hear that, I think, similar to how it's arguably we haven't fully adopted all the capabilities that Revit has to offer or BIM has to offer when we made the move from CAD to BIM. What is the downside right now, if I don't put any work into exploring this this new product or to exploring AI. How do we get people to set aside time, the time they need to really understand the opportunities of what AI can do for them.

B Well, a lot of us are saying that the best way to understand AI is to do AI. And that's really true. So I can explain till next week about what chat how to use ChatGPT, but I just recently introduced this to someone who had never used a chatbot, and they were writing an article about their retirement. And they said, well, this would be a good chance for me to write about my retirement, and I'll ask this chatbot to actually write it for me. And I'll just give it a few words, like, how long I've been doing it. I was a project manager, and I was this and that. And it wrote this amazing outcome that, was published in LinkedIn and got so much response and and all that. And and so if you think about this, this is what I feel drives creatives in a lot of ways is that the self satisfaction of producing things.

B And so there's so much enjoyment from this person having produced this work and getting their reaction that they had that they are more likely to use the system again. And so that that's how I think it works. And especially in the imagery part, I I follow all the major imagery. There's probably about 20 of them on LinkedIn who are just doing absolutely fantastic work. And I think the thing is is that we enjoy looking at it. They enjoy doing it, and they're satisfied by their outcome. So I think that's driving. So the way I would answer this that is to try to help people to just try it and see if they can do something that they feel really good about, and then let that lead them into something else.

B The other part about you know, you were comparing Revit to this is I mean, you have to go there are 3 major disruptions I was referring to. So hand drawing, CAD, and then Revit, and then now this. So this would be number 4. And, they all happen in my lifetime too, so it it's actually, gives me some some perspective in a lot of ways. But I think the uniqueness about AI that sets those other 3 apart is that I always think of AI as the imagination machine. And it has sort of this infinite potential. So if you look at a tool, say, for drafting or for creating database drafting, There's so many limitations that may seem fairly large, but in my view, the limits of AI and context of designers and architects is infinite. So there's so many things that can be created.

B I maybe I'll come up with an idea to help someone create something. I've heard people ask me for things, and I would say to them, call that company and ask them to make that. And they probably will, you know. That this is something this is actually an interesting point I find really necessary to discuss, is that, typically, over my career, tools for architects were things we had to buy. You know you know what I mean? We have to buy them. Now I saw in the eighties how a lot of software companies were hiring architects to work for them so they could have a better insight on their product development. And I'm sure that happened in the Revit age as well, But that's gonna have to happen in the AI age. In other words, if we want it to be a better tool for architecture, we need to get involved with the discussion with the people who do the development of the software, just not grabbing every little thing that we think we can find.

B It's sort of like if you think about the electric vehicle, I mean, there's a lot of things that people ask for that it wants to do. We want self driving cars. You know, there's things where I feel like the public is really engaged in trying to help that product develop. We need to do the same thing and not just let the products that are on the shelf at the grocery store happen, but how do we make the products on the shelf in the grocery store? How do we contribute to those products and then make them more useful inside our environment?

A I completely agree with you there in terms of being a part of the creating rather than just being handed the tools that we have to ultimately use. My question for you then, we have a lot of younger listeners on Practice Disrupted and a lot of young architects are new, new firm leaders. For the young leaders in architecture firms that have the power to bring these type of tools into practice, what advice do you have for them when it comes to investigating where where they should begin and and how they should look at in our integrating or using the tool?

B Yeah. It reminds me of the when I was 22 years old or 23 years old, and I said something to my boss like, wow, we should get a computer. And and it's sort of I got this strange look because, you know, the cost was enormous. And we had to build a computer room with computers the size of refrigerators. And I I actually here's an interesting thing that most people don't know about me is I had to work the night shift on the CAD machine because we had to run the CAD machine all day and all night in order 24 hours in order to pay for it. So I think to young people, what I say is that, a couple of things. First of all, I would recommend that we read everything. I always recommend this to students when they talk about resilience in their career.

B And one thing is is to not isolate themselves totally in the architecture, but understand the world. So looking at science or physics and, medicine and reading as much as you can to understand the whole landscape of information that has an effect on humanity. So I think that would be number 1. The second one is to spend a couple of hours a week doing something creative. So even though in my I have to be honest in the beginning of my career, yes, I did toilet rooms. I mean, I I will admit it here on on the radio. The the main the main thing is is that but I did very good toilet rooms. I mean, I got really good at it.

B It's something that I'm actually proud of. But while I was doing these things in the office, I always took time to be creative. I started painting, or I started doing assembly art, or I started doing jewelry kind of work with other schools or with art galleries and things like that. Creativity is something that individuals own. It's not something that someone gives you. Use that ability to go out and do other things that feed the muscle in your brain that makes that happen. And I I tell a lot of young folks this story about that. And I think it's really too that just it's not like I said, well, you gotta put in a 100 hours a week to do, you know, beyond the work hours to try to figure this out.

B The main thing, even if you spend a couple hours a week, reading about artwork in Italy or or maybe doing some sketching or doing mid journey for a couple hours a week. We could find a couple hours a week to do something like that, and I think that would be important for them.

A Right. And then I guess, what is your biggest hope for the future of architecture and technology when you look at even even the next 5 years and its and its potential evolution and all the opportunities that you talk about?

B Alright. So here's the landscape, and this is based on my conversations with the US Labor Department and also understanding the dynamics of population growth in the world. So by about 2050, there'll be 10,000,000,000 people. There's almost 8 now. There'll be 10,000,000,000 people on Earth. But if you think about that, about 2,000,000,000 people are gonna move from rural areas to cities. You see this in China. It's happening now.

B What that means is that we have to house all those folks. And when you take that number 2,000,000,000 over the next 25 years and you divide it by, let's just say, a 20 story building with 10 units per floor, that's 200 units. And if you look at that, I've done the calculation that, essentially, what it means is we would have to build a tower every every 30 seconds for the next 25 years, and we haven't even started doing it. So what that really means in terms of having a new tool like AI is that we might have the potential to make our work more efficient so that we can meet the demands of the public through architecture and and make just essentially make that happen. What I mean by the labor statistics perspective is that they see sometimes that the growth of the construction industry is limited because there is only so many architects and engineers that can make it happen. But what if I can make, in my firm, a 100 buildings a year? Just we'll just pick a number. What if I could use something like AI, and all of a sudden I can take the same number of people and do 300 buildings a year? And without any person working any more hours, maybe I'll give them all a raise, but essentially, that's the kind of efficiency that AI can do. And I I know this for example.

B I'll give an example of a person I work with at the university, and his specialty is coatings and paints. So if you think about it, their goal is to find the first and the most perfect coating on earth. It's the best paint you could possibly make. But paint is based on molecular science and chemistry. And if you think about all the combinations of chemical agents and molecules that you could think of and test, it it takes a long time to do that. But if he could only do a 1,000 a year, what if he could do 5,000 a year with the assistance of AI? He doesn't have to actually put in more effort, and maybe he can hire maybe just one more person. But, essentially, the impact of what AI could do for these scenarios is gonna help us meet the needs for a growing global population and also enhance the careers of the many people who are trying to make that happen.

A Hi, disruptors. Thank you for joining us today on an episode of Practice Disrupted. If you like the content for today's show, you can find all of our past episodes over on practice of architecture.com/podcast. Be a part of the conversation by joining me, our speakers, and other disruptors in our community at practice of architecture.com/community. Our social media handle is practice of arch. That's practice of a r c h. We'd love to hear from you, so feel free to drop us a DM and say hello. Tune in next week for a new conversation on change in the profession.