Can AI Revolutionize How We Design and Build with Wood?

Every week, more than 100 million people log on to ask OpenAI’s ChatGPT an estimated 1 billion questions, and in the same time frame its built-in companion text-to-image AI-powered app DALL-E 3 generates at least 14 million images based on user prompts. In just seconds, the former can write an article with the same title we’ve given this one (its actual usefulness, debatable) while the latter can render a skyscraper in the shape of a banana (its usefulness, also questionable).

But for all its meteoric trendiness, AI is no flash in the pan. It is poised to transform many aspects of industry in the coming years, including architecture and construction. And AEC sector players, big and small, are looking to it as the next frontier for developing practical applications that can solve real world design problems. So when it comes to architectural use cases in general, as well as those specific to wood construction, what are AI’s emerging opportunities for AEC professionals? And how can they be fully realized in the years and decades to come?

Banana Skyscrapper
Rendering Credit: ChatGPT | DALL-E 3

A Sustained Trend: Industry Is Investing in AI’s
Architectural Ascent

Among the architecture powerhouses exploring these questions is software giant Autodesk, which has made a significant investment in its very own AI Lab and published nearly 50 peer-reviewed research papers on the topic​. In late 2023, the company launched Autodesk AI, a new technology initiative that augments its AEC offerings with AI. Those features include machine learning and generative AI functionality integrated into its Fusion 3D CAD software, as well as AI features powering its Autodesk Forma urban planning tool.

Mike Haley, who leads the Autodesk Research group, sees a big upside for the sector. “Generative AI has incredible potential,” he writes in a recent post on the topic. “McKinsey estimates that it will deliver $2.6 to $4.4 trillion in productivity value. Things are going to get far crazier in the next 10 years, especially as generative design and generative AI create workflows that are faster, easier, and more accurate than ever.”

Autodesk is not alone. According to the firm’s 2024 State of Design & Make Report, the architecture, manufacturing, and media sectors are expected to boost spending on AI, with 77% of companies reporting that they will increase or strongly increase investments in such technologies over the next three years and 79% seeing AI yielding greater creativity.

Early Days: Embracing the Exploratory
Phase of AI

But grappling with AI’s potential impacts—both the advantages and the pitfalls—and determining how best to invest in its technologies and tools can be daunting for AEC firms, especially at this nascent stage.

“We don’t know where this is going,” says Phil Bernstein, an associate dean and senior lecturer at the Yale School of Architecture. “It’s clear [generative AI] provides a whole new avenue for generating ideas, and for doing things like prediction analysis and education. I think the healthiest attitude is the one we take about all tools, which is: It’s just one in your toolbox.” 

The architect, technologist, and author, who once served as senior vice president of strategic industry relations at Autodesk, is helping tackle some of the biggest questions facing the industry in his most recent book, Machine Learning: Architecture in the Age of Artificial Intelligence.

“It’s too early for anything other than wild speculation, but what I think will happen is that forces outside the profession—real estate development, manufacturing, supply chain management, construction management, finance—will likely create demands on architecture that we’ll have to respond to,” he says.

In his view, AI’s impacts on the sector will be profound, but are yet to shake out: “The advent of machine learning–based AI systems demands that our industry not just share toys, but build a new sandbox in which to play with them. Architects need a strategy for facing the opportunities and threats of these emergent capabilities or risk being left behind,” he says.

His biggest piece of advice when it comes to machine learning and generative AI is to embrace this early, exploratory phase. “It’s way too early to be making intergalactic declarations about what’s happening, because things are moving way too fast. Firms with resources need to have someone on their staff watching and building platforms for experimentation.”

Envisioning Possibilities: Imagining Wood Buildings
with Generative AI

Exploration is the approach Ramin Rezvani, a senior associate at Ankrom Moisan—an architecture, interiors, and urban design firm with offices in San Francisco, Portland, and Seattle, and a special interest in mass timber—is taking with his firm’s recent use of Midjourney. This text-to-image AI tool has been described as “ChatGPT for designers and artists” and has taken the architecture community by storm in recent months. 

From Rezvani’s view, a compelling place to start with AI is using it for rapid prototyping and idea generation, and he sees it particularly well-suited to mass timber. He’s explored a growing number of generative AI tools—Midjourney being one—that can help design teams envision the possibilities of materials such as mass timber during the early design and precursory phases of a project. Similar tools include DALL-E 3 and Adobe Firefly.

The wood-savvy firm recently used AI software to develop imagery for a mixed-use mass timber project it is currently designing.

“We’re really excited about mass timber as a low-carbon material and see generative AI as an effective tool to explore it conceptually with our clients,” Rezvani says. “In this case, what emerged was a series of unexpected iterations—happy accidents if you will—revealing novel ways to conceptualize mass timber as a structural element.”

“This process is driven by AI’s capability to combine vast amounts of data into new visual forms. And while it’s just idea generation at this point, it was effective in engaging clients, opening up discussions on innovative applications of mass timber in the project.”

But firms looking to incorporate AI into their practice must leave room for trial and error, Rezvani says. “I’ve been approaching it almost like a science experiment, changing the prompts slightly with each iteration to see what I get back visually with each update,” he says. “But it’s not like you can master it because it’s changing so rapidly. It’s a quick way to get the team on the same page and discover interesting emergent qualities from concept intersections that you may not have discovered on your own.”

Supercharging Wood Fabrication,
Construction, and Disassembly with AI and Robotics

Along with conceptual design, there are an increasing number of use cases where AI-powered technologies are helping in the fabrication and construction of wood buildings. 

One example is A.I. Timber, a new method for cross-laminated timber (CLT) production developed by the construction-technology company Maestro. The start-up firm is born out of the Turin, Italy-based firm Carlo Ratti Associati and co-founded by architect and MIT professor Carlo Ratti and senior designer and Cambridge grad Mykola Murashko.

“Produced by nature, with a little help from AI,” is how Murashko sums up A.I. Timber. The technology reduces waste by preserving the original shapes of the trees. Many lumber producers already use some AI and machine-based learning technologies to optimize the dimensional lumber output from a log. In the case of Maestro Technologies, instead of sawing unique logs into standardized boards, it uses proprietary AI software to scan each log and create irregular boards that fit together like puzzle pieces. 

“We’re connecting imaging technology with AI in order to generate custom machine instructions to a computer numerical control-powered sawmill,” Murashko says. “Our algorithm minimizes the loss associated with the volume of timber that needs to be cut in order to tessellate a set of lamellas [pieces of wood] into a CLT layer.”

“With a traditional computational algorithm, you would spend years, maybe millennia calculating what is the most efficient method, but with AI it’s infinitely faster,” he adds. The project’s first prototype debuted at the Digital Futures exhibition in Shanghai, China.

While A.I. Timber is still in the prototyping phase, Daisy is a start-up with AI software already being used by timber fabricators and manufacturers to create optimal structural wood-frame designs. 

“It’s a form of a generative AI, but it’s taking it to the next level, beyond conceptual design,” says Luke Whale, technical director for the company. “It’s looking at actual structural design and the layout of componentry. It’s a detailed design tool that completes a structural analysis to code and will come up with a framing layout and cost which is absolutely bang on—optimized for cost and other criteria.”

Using Daisy AI, a building professional can upload an empty floor layout and download a full, production-ready light-frame structural floor design in less than half an hour. The designs include a summary of all costs and a components file including joist coordinates and decking sheets.

More recently the company collaborated with Autodesk to explore how both AI-powered applications could be used together to speed up wood-frame floor and wall designs, while reducing the amount of concrete used in foundations—thereby reducing a building’s total carbon footprint.   

When it comes to the future of AI and timber construction, Whale is cautiously optimistic. “Inevitably AI is going to have an impact on the design profession and the way the workforce operates,” he says. “This software opens up opportunities for those maybe less familiar with timber construction. But in most cases we are reducing the time-consuming grunt work—and in my experience most designers’ in-tray is chronically bigger than their out-tray—so this is freeing them up to do the more complex, innovative thinking that AI can’t do.”

Generative AI and robotics are also entering the fray when it comes to more cost-effective prefabricated home construction. Autodesk has collaborated with California-based Factory_OS to enhance the design and production of modular light-frame wood housing units using machine learning. 

“What started out as a concept two years ago has now developed into a full-force effort to out-design carbon emissions while providing quality, affordable homes for everyone,” says Factory_OS CEO Larry Pace in a blog post.

The Citibank, Autodesk, and Google-backed manufacturer is exploring how AI and robotics can help build lower-carbon, wood-frame modular units for single family homes and larger multifamily projects cheaper and faster.

“The impossible problem that maybe AI can help with is how are we going to drastically increase the total amount of floor area while we’re drastically decreasing the total carbon emissions from all buildings?” asks David Benjamin, director of architecture, engineering, and design research at Autodesk, in a blog post for Slate Technologies.

Designing a new affordable housing development with generative software can eliminate the need to repeatedly redraw the prefabricated light-frame modules to prioritize various factors, such as carbon emissions or cost. “An architect could optimize for specific attributes, such as heating or privacy,” Benjamin explains, “while the system can explore hundreds of options really quickly.”

At a Bay Area start-up located not far from Factory_OS, AI and robots are also streamlining how wood buildings can be deconstructed for potential re-use. “We create robots that reclaim wood,” says Jorie Wisnefski, the marketing manager for Urban Machine

The concept sounds simple, the execution is anything but, she explains. The San Francisco-based start-up is reinventing how wood is salvaged and reclaimed using complex AI-powered robotics, in novel ways never done before. This includes using AI-powered cameras and X-ray inspection to assess the wood’s condition and suitability for re-use, including removal of any metal fasteners or fragments. 

“Urban Machine has been testing the robot in real-world scenarios for the past two years, constantly improving the design using advanced machine learning,” Wisnefski says. “Through the use of AI, the robots are trained to recognize a growing catalog of metal fasteners.”  

The technology offers a promising way to extend the life cycle of wood products, further enhancing the material’s carbon footprint.

Future Outlook: High-Tech Meets High-Touch

AI undoubtedly has the potential to transform wood architecture, along with the broader design world. But in this burgeoning, exploratory stage, it’s unlikely we will see exponential efficiencies or seismic shifts in how the AEC industry does business—at least not yet.  

The good news is AI can take on repetitive tasks and less mentally stimulating work, freeing up design teams to devote more time to higher-level problem solving and design challenges. 

And while AI may be perfectly suited for making the AEC sector more productive, the real opportunities, from Bernstein’s perspective, may be in rethinking the industry at large—and reinforcing the enduring value architects can bring to their clients. 

“Architects need to clearly define their value proposition in an era of computational automation to avoid being marginalized by AI,” Bernstein says. “The integration of AI will not just be about efficiency, but also about creating new systems from which design can spring, leading to meta-design processes.”

He adds: “The true value of AI in architecture will be realized when we use it to enhance the quality of the built environment, while maintaining the human touch in design.”​ ​

Back to top

Get wood innovation
in your inbox.

Sign Up!