Jun Cen

“Architecture was more of a handcrafted object, like [during] the Renaissance, when I first emerged,” said the late architectural pioneer Beverly Willis, 95, during a recent phone call as she discussed the impacts of technology during her long career. “And then it blossomed into this business, this science: the science of the soil, of the air, of things we didn’t consider when we started.”

The big pivot during Willis’s career was the adoption of technology into architectural workflows, a shift that holds lessons for the ways artificial intelligence could reshape the profession today. In the early 1970s, Willis spearheaded the development of Project CARLA (Computerized Approach to Residential Land Analysis), an early digital tool that allowed for relatively rapid site evaluations for large-scale projects.

At the time, architects were being tasked with designing and planning larger and larger developments, adding complexity that was challenging to navigate with existing tools. CARLA helped devise a plan for Aliamanu, a 525-building housing community in Honolulu near Pearl Harbor, Hawaii, that required significantly less earth-moving than similar projects, shrinking the development footprint and lowering construction costs. Willis saw the potential immediately: The program condensed hours of work into mere minutes.

The experiences Willis had decades ago with Project CARLA mirror architecture’s shift toward AI today and offer still-relevant insights into the ways technological sea changes challenge preconcieved notions of creativity, labor, and business practices.

Architectural theorists and practitioners are divided over visions of the future of their work. Do new technologies have the potential to open opportunities for more creativity and community outreach, or do they devalue the work of architects and further stratify an economically precarious industry?

Everyone agrees on the inevitability of change. It’s only a matter of time, says Daniel Koehler, an architect and professor at the University of Texas at Austin who studies AI, before programs will be able to create drawings and blueprints based on a designer’s prompt that comply with code restrictions and result in workable blueprints.

These predictions have fueled a rush toward AI by established tech design players like Autodesk, which has invested in generative design startups, while at the same time inspiring trepidation and anxiety from architects over their future roles in this design landscape.

“You can see [AI] as being fantastic or consider it terrifying, but it’s going to radically change everything, and we’ve just started,” says Neil Leach, a British architect and theorist with a more pessimistic view of these tools. “Architects aren’t engaging with the bigger picture. They’re ostriches with their heads in the sand.”

For decades, architects have leveraged technological advances to realize their visions and gain new opportunities and commissions. Peter Eisenman’s 1987 Biocenter design for J.W. Goethe University in Germany included computer-generated-design outputs that showcased the potential of computer-aided design, and Gehry Partners created its own digital design interface, adapted from aeronautical technology, to realize the starchitect’s swooping forms.

Bevan Bloemendaal, chief creative officer at Atlanta-based Nelson Worldwide who oversees teams focused on retail design and architecture, says he has observed excitement from architects at being pushed cognitively and creatively by these tools, but also fear about losing their jobs. He’s had his employees experiment with AI by placing the programs through design sprints and has found that his creative staff are seeking out opportunities to learn and embrace these tools. Many big-name firms, such as the London-based Heatherwick Studio, have already noted their use of AI for pre-planning and design concepts.

Amid the demands of a tool that has the potential to both liberate architects from busywork and threaten job security, Bloemendaal is most concerned about maintaining human-centered design.

“We’re businesspeople. [We] want to do more and [do it] better,” he says. “Do we want the client to get the most for their dollar? Of course we do.”

AI will save time and labor costs, argues architect, author, technologist, and Yale professor Philip Bernstein, FAIA, especially when it comes to image generation and technical production, as well as data and text management. But as with previous technological disruptions, the key will be how firms decide to utilize freed resources. He suggests three scenarios: investing in refining a design and finding better solutions to design problems; improving services through predictive and performative delivery and breaking away from commoditized fee structures; or lowering commoditized fees.

“If history is a guide, the profession will find it hard to resist the allure of the third scenario, particularly in times of economic stress,” Bernstein says.

Others believe these resources will be reallocated in ways that improve the profession. Adam Tank is co-founder of Transcend, a Princeton, N.J.-based software firm that creates generative design tools specifically for infrastructure design and construction. Autodesk recently invested in a $20 million Series B round of funding for the firm, part of the company’s string of big acquisitions and investments in AI, including a $240 million buyout of Spacemaker, a generative design planning tool relaunched as Forma.

Since AI software can do all the repetitive work, such as making calculations or reformatting Word docs, that eats up the average workday of an engineer or architect, firms utilizing Transcend have been altering their work processes, Tank says. Designed by engineers and architects, Transcend spits out plans and designs complete with engineering calculations, allowing humans to double-check plans before anything gets built.

Tank points to widespread reports of talent shortages in the industry—the U.S. Bureau of Labor Statistics predicts the architecture profession will grow at a rate of 5% per year through 2032, with more than 8,000 annual openings, and also forecasts an annual shortage of 25,000 civil engineers—as a sign that AI will enable more productivity instead of cutting jobs. He’s observed firms moving existing engineers toward detailed design and project execution, instead of investing their time in preliminary design work or budgets. There’s also more room for what could be called sales and marketing work; teams, including architects, are able to spend more time on detailed bids, differentiating themselves from other firms.

“[With AI,] architects can focus on what’s more energy efficient, or what has the smallest impact on the surrounding community,” Tank says. “That’s how people will be spending their time. That’s the future."

But being able to do more for less and take a larger portion of the overall pie for human employees doesn’t mean the pie gets any larger. British architect Leach predicts that marquee firms, like those led by starchitects, will likely still charge premiums relative to the competition. Suddenly, they’ll be able to take on more work and do it with fewer people, becoming what Leach calls “superusers” who can afford more cutting-edge, powerful tools.

“As far as design generation, let’s hope that the ability to have machines generate ‘first ideas’ doesn’t create a race to compete for work by presenting more and more elaborate work in interviews,” Bernstein says. “This is an approach that I think may be inevitable but still constitutes—whether AI-inspired or not—giving work away for free.”

The nascent union movement in the architecture profession has also noted the potential impact of AI. When asked about technology’s impact, the union at Brooklyn, N.Y.-based Bernheimer Architecture sent a statement noting that the union and management “understand the potential for artificial intelligence to be integrated with design tools in innovative ways. AI technology, however, should not take the space and place of human labor.” In addition, the statement read, “This work requires an understanding of human relationships, physical space, and emotional responses ... something automated AI processes can never replicate.”

For Leach, the most vital calculation around the impact of AI is labor and productivity. Architectural offices tend to get paid as a percentage of building costs, roughly 5% or so in the U.S. and Europe. He expects that when AI becomes more ingrained in studios, and firms have the ability to do more with less, they will be able to undercut that fee, which will lead to a race to the bottom. He says that in China, fees already have approached 2% due to more widespread use of these technologies.

In a few years, he believes there will be a single platform that can take design and construction from data through to fabrication, evaluating structural, regulatory, and environmental factors along the way. He points to Xkool, a Chinese architecture program that has invested significant time in the tool’s ability to label images to better serve the needs of architects.

“What we need to design right now is not another building, but the future of our profession,” Leach says. “We need to think of new ways of operating.”

“There’s a fork in the proverbial value-proposition road here,” Bernstein says. “The firms that use the resulting efficiencies to lower fees will drive down prices for everyone; the firms that use the technology to deliver greater value will make more money. I hope for the former and worry about the latter.”

The question facing architects is how these tools get developed and how industry economics grow alongside technological evolution. Some fear not only stratification but, eventually, the potential for certain projects to simply sidestep the use of architects altogether.

Willis admits that throughout her lengthy career, she sometimes remained skeptical of the technology she pioneered and didn’t always shift her firm’s focus toward tapping into its potential. But she did believe that technology was a great enabler.

“It helps you be more creative, because you understand things and can measure things that all help you make good decisions,” she said.