In December, the American Institute of Architects published Artificial Intelligence Adoption in Architecture Firms: Opportunities & Risks as part of their The Architect’s Journey to Specification series. The growing use of AI in architecture has prompted an examination into the perception and goals of these tools for firms and organizations, and what a future of AI and architecture might look like. Though this report aims to provide key insights about the profession’s current use and perception of AI, its methodology has glaring issues. Out of 10,000 randomly selected AIA contacts, only 541 completed the survey. Additionally, a 51% majority of respondents were in leadership positions as firm partners and principals. The report, mirroring the profession, also lacked gender diversity: only 25% of the respondents were women. In this way, those who voluntarily responded to the survey are not representative of the professionals who may be asked to use AI in their work. Only 3% of respondents, roughly 17 people, were non-licensed design professionals. This demographic likely includes Associate AIA members early in their career, including myself, who may be most affected by the future of artificial intelligence in the architecture and design profession. Firm leaders may also be driven by practice management targets that value efficiency and cost, potentially skewing the data in favor of AI usage. Respondents were optimistic about AI streamlining and automating tasks, with firm leaders responding even more optimistically than project architects. Notably, the report was published in collaboration with two project management software companies, one of which advertises AI-assisted software offerings.¹
Despite a flawed data set, there were a handful of insights that may be more widely applicable to the profession. According to the report:
Data graphic from the AIA’s report Artificial Intelligence Adoption in Architecture Firms: Opportunities & Risks. March 11, 2025.¹
While 75% of respondents expressed a shared curiosity and optimism about the use of AI in the profession, it is clear that concerns about its use are even more widespread. The report also found that only 6% of respondents use AI regularly for their jobs, and 82% of architects are interested in a charter for responsible use of AI in architecture. Interestingly, the report found that AI experimentation is driven by architects aged 50 or younger, and these architects are also significantly more concerned about AI than their older colleagues.¹
In addressing these widespread concerns, larger firms might enhance security with proprietary software, but this requires intensive resources and training. In the meantime, AI chatbots and image generators are trained on user prompts, posing unintentional consequences for design integrity and confidentiality. AI generated images may also lead to a circular design phenomenon, where AI tools learn from designers’ input, and designers subsequently seek inspiration from AI tools. One creative aspect of AI tools, the art and science of generative prompt writing, has promptly been farmed out to chatbots as well.
Architecture can become easily insulated, but we are often designing and collaborating with individuals outside of the profession. On April 3, the Pew Research Center published a report titled How the U.S. Public and AI Experts View Artificial Intelligence. Compared to the AIA’s report, this report sampled 5,410 adult respondents from the Center’s comprehensive American Trends Panel, and 1,013 AI experts. This report highlights the difference in perception about AI use between the public and AI experts, with experts being more generally optimistic about AI than the public. The data also revealed that men are more optimistic about AI than women, especially among experts where “63% of men say AI’s impact on the U.S. over the next two decades will be very or somewhat positive, compared with 36% of women.” AI experts and the public seem to agree on two major factors: both groups want more personal control of AI, and worry that the U.S. government will not go far enough in regulating AI use.² The results of this report pose another interesting question: how do architects weigh public perception of AI with that of experts when deciding how to use AI tools? Clients are rarely AI experts themselves, and will a client lose trust in their architect if an AI generated image with inaccuracies is presented? Image generators cannot contend with larger design contexts, such as site conditions, adjacencies, ecosystem, culture, community, zoning limitations, end-user, and constructability. Architects are retained for this expertise and knowledge, and AI generated images may provide unrealistic or inaccurate design concepts. With this in mind, “NCARB’s Position on the Use of Artificial Intelligence in the Architectural Profession” was published in October 2024.³ This report was created by 150 licensing board members at NCARB’s Annual Business Meeting (ABM), who seemingly had difficulty reaching a consensus on AI use as well:
Excerpt from NCARB’s Position on the Use of Artificial Intelligence in the Architectural Profession. October 17, 2024.³
ARE buzzwords litter the statement, but these concepts are crucial as architects move forward with new AI tools in their work. If AI is inevitable, then AI ethics and standards need to be discussed in tandem.
As always, there will be architects who continue to take a very hands-on, analog approach to design. Students will study the computer-free collages of Tatiana Bilbao, Sou Fujimoto’s sculptural potato chips, Frank Gehry’s crumpled paper buildings, and Jennifer Bonner’s dollhouses.⁴ ⁵ ⁶ As a student, I was drawn to classes that indulged representation methods as a means of design and mentors that encouraged radical approaches: graffiti can be a rendering, a model can be a book, a one-point perspective can be a model, a turntable can be a nightclub. In these studios, creating images and diagrams became part of the art of architecture. In contrast, the image tone and quality of “archviz” out of rendering software has become recognizable, even to AI image generators.⁷ Sharp, high-contrast, impossibly lit images with one protruding building corner and a figure who is inexplicably in two places at once. AI image generators have become known for their inaccurate depictions of people, placing limbs in the wrong location, distracting facial expressions, too many fingers, strange cadence. Architects’ retrospective exhibits may be filled with AI imagery in the near future, moody renderings with an uncanny entourage. Design of the built environment is inherently social, and at minimum architectural visualizations need to reflect a believable, or at least undistracting, end user. If narrative and images are relegated to AI tools, we may lose an essential human touch in architectural representation. It's clear that more research needs to be done on the use of AI tools in the architectural profession, both individually and collectively. The available data confirms that larger firms are spearheading the use of AI, potentially running ahead of its consequences and widespread concern. Maybe I’ll be left behind with a potato chip model.
Model from Sou Fujimoto Architects’ “Architecture is Everywhere” Installation, Chicago Architecture Biennial 2015.⁵
Aryana Leland is a human designer based in Denver, Colorado. She studied Architecture + Art History at Cal Poly Pomona, and her current interests are Pikmin and postmodernism
“Artificial Intelligence Adoption in Architecture Firms: Opportunities & Risks.” The American Institute of Architects, December 2024.
McClain, Colleen, Brian Kennedy, Jeffrey Gottfried, Monica Anderson, and Giancarlo Pasquini. “How the U.S. Public and AI Experts View Artificial Intelligence.” Pew Research Center, April 3, 2025.
“NCARB’s Position on the Use of Artificial Intelligence in the Architectural Profession.” NCARB, October 17, 2024.
Frearson, Amy. “Tatiana Bilbao Says She Banned Renderings in Exclusive Interview.” Dezeen, December 4, 2019.
McKnight, Jenna. “Sou Fujimoto Creates Installation from Potato Chips and Ping-Pong Balls.” Dezeen, October 23, 2015.
Bonner, Jennifer. “The Dollhaus.” MALL, 2017.
Iontcheva, Ina. “What Is Archviz? Everything You Should Know.” Chaos, February 21, 2024.
Zewe, Adam. “Explained: Generative AI’s Environmental Impact.” MIT News, January 17, 2025.