The Future is Now: AI Transforms Archviz – What's Next?

Welcome to the Creative Viz podcast, where we dive into topics on architecture, development and visual design. Today we’re tackling a transformative topic in our field: the integration of AI and real-time rendering. Generative AI is top of mind for many of us in creative industries, and I have actually held off on broaching the subject here for a little while as it has been evolving too quickly.

A year on from the introduction of ChatGPT 3 and the wider adoption of generative AI tools like Midjourney and Stable Diffusion I believe we have a better sense of where things are headed now and hopefully we can start to make sense of it all. But by no means will this be the last word on AI & archviz - stay tuned as I’m sure there are new surprises just around the corner.

For now, let’s explore how these technologies are reshaping our industry, from enhancing creativity to revolutionizing workflows. We’ll start with the promise of AI, then explore some of its limitations and go on to discuss important ethical considerations as well. I’ll weigh in on whether or not AI will replace you, and finally gaze into the crystal ball to make some guesses as to where this could be heading.

So… in the world of Archviz (architectural visualization), AI and real-time rendering are more than just emerging technologies; they represent a paradigm shift. Gone are the days of static, time-consuming rendering processes. Today, AI’s infusion in our field is leading to groundbreaking changes. Real-time rendering, powered by advanced AI algorithms, is not only speeding up our workflows but also allowing us to experiment with designs in ways we never thought possible. It’s an exciting time for architects and visualizers, as we stand on the brink of a new era of digital creativity.

What does this mean for the traditional methods we’ve relied on for so long? Well, it's not just about replacing the old with the new, but rather enhancing and augmenting our existing skills with these powerful tools. AI and real-time rendering are enabling us to push the boundaries of architectural visualization, offering a glimpse into a future where our ideas can be visualized almost as quickly as we conceive them.

AI is already assisting in complex tasks like lighting adjustments, texture creation, and even generating entire scenes autonomously. Real-time rendering, once considered to be the holy grail of rendering, is now a daily reality for many of us, offering immediate visual feedback and drastically reducing the time from concept to visual representation.

For designers, these advancements are not just about speed; they're about the opportunity to explore and ideate like never before. With real-time rendering, iterations that once took days can now be explored in hours, if not minutes. This rapid turnaround is changing the way we approach design decisions, allowing for a more iterative and dynamic creative process. It’s a liberating shift, freeing us from the constraints of time-intensive rendering and opening a world of creative possibilities.

For visual artists, real-time rendering can remove much of the tedium associated with creating realistic textures and the imperfections that make a digitally-created image look more photographic. Entourage - that is people, trees, and cars in architectural scenes - benefit in particular from the ability to use generative content instead of relying solely on 3d content. It doesn’t work every time, but replacing 3d “props” with generative substitutes can substantially speed up the process of adding entourage to a still image. Additionally, there is a notorious lack of diversity in existing entourage libraries, and AI algorithms can do a much better job of creating figures that bring a sense of realism and scale to the scene.

Having said all of that, there are significant limits to what is possible today and what will likely be possible for some time with AI content. When you look at an architectural rendering, many of the textures and almost all of the entourage in the scene are stand-ins. That is, they are important elements but could be swapped out for a substitute very easily with little or no effect on the scene as a whole. This is where AI really shines.

Most of the fixed elements in the scene however, are truly custom one-off pieces. No building is the same as another, and very few interiors spaces are the same - if they were, then you wouldn’t need a rendering? So the architectural spaces themselves need to be invented every time with exact fidelity to the design documents. In additional to the building itself, many elements within the space are unique as well. There are often custom furniture, art pieces, and objects specifically designed to inhabit the space to be visualized. These all must be modeled in 3D and assembled according to the architectural and interior design.

Another thing to consider here is that the location of the new building in its context is usually a crucial part of the messaging. For an exterior rendering in an urban setting, the existing adjacent buildings often need to be rendered with as much care as the new one in order to give a faithful representation. Early-phase development will be granted approval or not based on how well the new building is perceived to confirm to the guidelines and requirements of the particular site.

For interior renderings, the view outside is often an important part of the story - consider the value of the view from a penthouse apartment. Although there are certainly times when a generic stand-in created by AI will do, in nearly all circumstances, depicting the view as accurately as possible is essential to the authenticity of the image. A lot can depend on this in a marketing campaign for instance. 

In both of these examples, this unique context must be created in 3D or with photography, or sometimes a combination of the two. AI is simply not up to this challenge - at least not yet.

Rendered archviz images need to be accurate of course, but they also need to be visually pleasing if not beautiful. There are existing algorithms that can go some way towards recognizing or even measuring beauty, but it is still up to the artist to compose the image properly. At best, the working with AI tools to generate a beautiful image can mean sifting through lots of options that need to curated and judged for quality. It remains to be seen if AI can actually speed up the process of setting up and composing an image that a viewer would find pleasing.

Then there is the problem of AI generated content being self-referential. The current set of AI tools are trained on existing content to generate something “new”. Most of these libraries are cut off in early 2022, just prior to the launch of the tool itself. As a result, we cannot expect AI to give us something we haven’t already seen before in many ways. Already, there is a certain look to AI generated images and it’s incumbent on the artist to not rely too heavily on AI in order to avoid creating something stale and lifeless. Therefore, true innovation in image-making is still very much up to the artist.

I also want to touch on the important ethical questions surrounding the use of AI imagery, especially around the authenticity and ownership of these visuals. As we integrate AI into our workflows, we must be mindful of ethical implications. We need to consider issues like copyright, the authenticity of representation, and the potential for AI to replicate existing designs or elements without proper attribution.

It’s crucial for us as professionals to stay informed and conscientious about these issues. We must establish and adhere to ethical guidelines that respect intellectual property and the integrity of our work. This responsibility is shared among developers, users, and the broader creative community. As we navigate this new terrain, let’s commit to using these powerful tools in a way that respects the artistic and intellectual labor that goes into architectural and visual design.

There are of course broader implications here well beyond archviz. I am hopeful that the AI algorithms will not get ahead of the creative process and that the creators of original content will continue to be fairly compensated for their work. For now, it’s important to be aware of the potential for this problem and be diligent about knowing where the content that you are using is coming from.

The integration of AI in Archviz also brings up questions about staffing and production efficiency. There's a common concern that AI might replace human jobs, but in my experience, it's more about shifting the nature of our work. AI can handle repetitive, time-consuming tasks, freeing up human talent to focus on more creative, high-level aspects of visualization. It’s not about reducing staff but empowering them to be more creative and strategic.

On this front, I am much more hopeful, as I believe AI can lead to more job opportunities in our field. As we can produce higher quality work more efficiently, the demand for skilled professionals who can leverage these tools creatively is likely to increase. The key is to adapt and evolve, to learn how to work alongside these technologies, harnessing their potential to enhance our skill set and the unique services that we offer.

So let’s take stock of where we are right now. For artists, if you haven’t done it already go create a free account, dive in and start making things. If you are used to creating images in 3D, you may astounded by the responsiveness and speed with which AI images can be created from scratch. You may also be astounded by the visual mistakes these tools can make. But dig in a little deeper and investigate ways to integrate AI into your existing workflow, especially for adding entourage and props.

For a client looking to commission visual design, it’s a good idea to be aware of the possibilities that these tools provide. But, be wary of any solutions that claim to automate the entire process. Ensure ethical sourcing of AI content and maintain open communication with artists about how AI impacts their production.

Looking ahead, the possibilities for AI and real-time rendering in Archviz are boundless. We’re already seeing the beginnings of what these technologies can do – from creating immersive virtual environments to aiding in complex design decisions. But we are clearly just getting started. In time, we could see a complete transformation of how we conceptualize, visualize, and interact with architectural spaces.

Imagine a future where virtual and physical spaces are seamlessly integrated, where clients can step into and experience a design in real-time, making changes on the fly that are instantly visualized. This level of interactivity and immersion could redefine the entire architectural process, from ideation to presentation. It’s an exciting prospect, one that promises to elevate our industry to new heights of creativity and innovation.

As we wrap up, it's clear that AI and real-time rendering are more than just tools – they are the first steps into a new era in architectural visualization. They offer us a glimpse into a future where our creativity is limited only by our imagination. So, let's embrace these technologies, explore their potential, and continue to push the boundaries of what’s possible in Archviz.

As I mentioned, we're going to continue to talk about AI developments in future episodes, but for now we’ll leave it there. I hope I’ve given you a lot of think about on the very important and timely topic of artificial intelligence.

Thank you so much for listening. Be sure to give us a follow to stay on top of all the latest developments and I will see you next time.