arch survive AI image

How the Practice of Architecture Survives Artificial Intelligence

There is nothing more human than architecture. It is created, made, and used by us. And yet for the past century, too much of architecture had evolved into a found object, a singularity, created in the black box of genius. The creeping cultural irrelevance of “Architecture,” caused in part by its insular culture and an internet obsessed with 2D images, may now have an even graver threat: artificial intelligence. AI threatens to eliminate “the middleman” in building, which in many cases is the architect. 

As with the Industrial Revolution, AI is not only inevitable, but it has the potential to create a better world for more people. So architects have a choice: become niche Luddites and dilettantes, or change the profession and use AI to enhance what they design. The choice is a stark one: Either we make change happen, or there will be quicker and cheaper alternatives replacing us. 

It is easy (and boring) to say things must change; it is harder to say exactly what needs to change. The training of architects, how they practice, and how they’re connected to the building process—all of this will change, whether architects want it to or not. So rather than rail against change, I will posit some directions that might mesh with the changes that we cannot fully control.

(1) TRAINING: Going beyond technology and aesthetics. The CAD monkey (and the degrees and training based on those skills) will soon become obsolete. The tools will be at everyone’s disposal. As a result, education will need to focus on the humanity of seeing, responding, listening, and adapting. AI is a nearly infinite database, but the database itself is rigid and inhuman. No software can simulate the human dimension of design. The humanity of building is often lost in architecture, and may cease with AI. So the way designers are trained must evolve beyond the studio model to embrace an understanding of building and the communication skills that are required before designs can be executed. At the very least, every academic project must have a nonarchitect client and a builder strategically present at reviews. 

(2) APPRENTICESHIP: The end of the intern. After AI inevitably revamps architectural service, the old internship model will die. Since the explosion of CAD, interns have in most offices become attached to the computer. That model ends when the design architect can input the design directly into the computer and use software that will resolve all the integration that is currently left to interns and young, unlicensed professionals. As a result, apprenticeship needs to be integrated directly into school curricula, as some schools now do, and it must be extended into those years before licensing, allowing for increasing levels of responsibility and greater understanding of how building happens, beyond the necessary drawings. This means that a new, intense form of “job shadowing” will be needed for those who were once chained to their workstations. These apprentices will learn by working with architects, builders, and clients, not by being isolated behind a monitor—IA will free offices of that need. What we do with that freedom will largely determine our value and continued relevance.

 (3) PRACTICE: Communication first, design second. Before AI, the academy lauded “genius” in architecture. It was held up as a model and a legitimate aspiration. With AI, need and solution will become one, replacing genius. Ironically, the use of AI by architects may be the only way to save the humanity of architecture, but only if architects can communicate their value. AI could enable architects to be the human face of building, but our object-focused training and practice need to be expanded to recognize the methods of creation, the human interactions of understanding and invention that software can only simulate. Intelligent design will allow builders to take any site and all relevant codes, overlay a “style” preference, provide a budget, and quite easily hang the “No Architects Need Apply” sign. But zoning issues are human issues; cost control and value engineering involve human values; construction options have aesthetic and maintenance consequences that artificial intelligence cannot register. Architects and those who build and use buildings are humans. That humanity is the advantage architects have over AI, our superpower, but how our humanity helps the process involves listening, collaboration, and building savvy, not the object obsession of recent generations.

(4) BUILDING: Process now, product later. Even with the emergence of AI, most buildings will need designs that mediate between the builder, the users, the community, and the aesthetics. Most buildings are built without architects now, but even more will be designed by AI services alone. Unless architects can offer the human-to-human understanding and communication that no computer can mimic, the digital process will come to exclude them as unnecessary cost and time dumps. Architects need to base their practice on listening, communicating, and partnering with clients.

Design is not enough to justify architects in the new world of AI. Architects, like lawyers and doctors, have long used the power of restricted knowledge to leverage value. But for architects, AI will create the WebMD or Legal ZOOM equivalent of design practice. The waves of technological change—which have been accelerated by the pandemic—are around us now. Do we sail these waters, or drown?



Get smart and engaging news and commentary from architecture and design’s leading minds.

Donate to, a Not-For-Profit website dedicated to reconnecting architecture and design to the public.