Designing with and Against AI: Voices from the Built Environment

October 23, 2026
BY Rawad nahhas

Somewhere between the utopian promises and the existential panic, architects, planners, and designers are quietly figuring out what AI actually means for the work of shaping cities and spaces. Across Tulane School of Architecture and Built Environment (TuSABE), faculty are grappling with the same question: what does it mean to work alongside a tool this powerful?

Meghan Saas, Director of the Design program at TuSABE, shared that “AI should serve as a collaborative tool rather than the beginning or end of the creative process,” emphasizing the importance of human thinking and questioning, something AI cannot replicate “the machine gets in the way of the making process.” She discussed her cautious approach to allowing AI use in the classroom, particularly using platforms trained on opt-in artist materials. Saas also addressed concerns about AI’s impact on ownership issues, highlighting the murkiness of intellectual property laws and the importance of teaching students to think critically about AI’s limitations, including its tendency to amplify existing biases.

Assia Crawford, Assistant Professor in Architecture, sees AI as a mirror, and a skewed one:

AI is very good at producing things that exist that are widely documented, and what we widely document is very skewed

She adds that who’s producing the information and what kind of information actually gets used in terms of these datasets is very much curated by a tiny portion of society. Crawford raises a serious concern about cognitive dependency. She points to studies suggesting that when we outsource mental tasks to AI, the associated brain functions begin to disengage. The quality of AI output, she argues, is entirely determined by the quality of the questions you bring to it, which requires deep specialist knowledge, not just surface-level familiarity. She is skeptical of AI as a design generator. She describes it as a “backward-looking tool.” It references precedents without cultural specificity or contextual awareness, blending Egyptian, Greek, Gothic, and Renaissance architecture without distinguishing them. Crawford also challenges outdated statistics on AI’s water consumption, noting that the widely-cited figures come from far less efficient models. Her graphic novel, “BIOS in Search of Zoe”, used the equivalent of 3–6 toilet flushes of water. That said, she acknowledges the net environmental cost is real.

John Renne, Director of the Real Estate Development Program at TuSABE, speaks on AI’s immediate impact on real estate and urban planning. His most concrete example is uploading entire zoning codes into an LLM to run regulatory scenarios. What once required specialized attorneys is now accessible to any practitioner. He also reframes AI as a neutral tool. In Transit-Oriented Development contexts, AI helps pursue multi-modal optimization (transit, walking, cycling) to reduce car dependency, but the values guiding that work remain a human responsibility. Renne touches on his work in urban resilience and AI: “Building AI literacy gives me better tools to pursue the resilience work, and the resilience work keeps reminding me what the tools are ultimately for.”

In class, Renne requires disclosure of AI use and holds students to their own standards of judgment. He’s also co-developing OralExam.AI with a colleague, using AI to scale oral/Socratic exams, which are harder to fake than written work, as a response to authorship concerns.

Ira Concepcion, Visiting Assistant Professor at TuSABE, finds that architecture has always absorbed technological shifts: hand drafting to CAD to BIM to AI. Concepcion frames AI not as a rupture but as a continuation of that evolution. Counterintuitively, Concepcion argues that large firms face more friction adopting AI due to bureaucracy and standardization. Solo practitioners and small firms are more agile, and AI gives them capabilities previously only available to larger offices. The danger, in his view, is firms attempting to algorithmize the design process itself. Design requires synthesizing intangible inputs: client conversations, site observations, personal instinct. That cannot be reduced to data. On a notably contrarian take, Concepcion argues the cost-benefit of AI’s energy consumption will ultimately favor environmental outcomes: “with anything that we build — even buildings — the first buildings we built were very inefficient. But we had to go through that iteration to get the highest performance buildings that we have today… we’re just in the infancy of this technology.”

What these four voices share is a refusal of the binary. None of them are dismissing AI; none are surrendering to it. They are, instead, doing the harder work of determining what to protect. The built environment has always been shaped by the tools available to those who build it. The question AI forces, is which humans get to hold the tools, and what they are asked to give up in return.

Related

Assia Crawford co-edits book titled Designing Mycelium

Assia Crawford, assistant professor in architecture, has co-edited a new book on one of the most unexpected materials in contemporary design: fungi. Designing Mycelium: Exploring the Design Potential of Fungi, published by Routledge, was co-edited with Jonathan Dessi-Olive and brings together voices from across disciplines.

The Review Book Cover

School releases The ReView, 4th Edition: Emergency

The ReView, 4rd Edition: Emergency, the school's annual publication featuring the work of students and faculty from the 2024-25 academic year, was released in September. To celebrate the book launch, programming included an invited lecture by noted architect Jing Liu, followed by a discussion with faculty book contributors Assia Crawford, Sonsoles Vela Navarro, and Shawna Meyer.