Critical Making meets AI
Helping the next-generation develop the intellectual tools to face design’s biggest challenge yet
By Ugonna Ohakim & Sayjel V. Patel
From Critical Making to Future Building
A s 2024 comes to an end, we find ourselves at a pivotal moment in the history of AI and design. Against this backdrop, we were invited to teach at the University of Waterloo. Throughout the fall semester, we met weekly with a group of architecture students, encouraging them to view AI not as something distant or daunting, but as a new design material — one capable of shaping learning, personal interests, and professional education in new and unexpected ways.
Our teaching drew from the philosophy of “constructivism” where learners actively construct their own knowledge (Papert, 1980). This aligns closely with Matt Ratto’s (2011) concept of “critical making”, which promotes design as a means to critique ideas and challenging social issues.
Our teaching drew from the philosophy of “constructivism” where learners actively construct their own knowledge (Papert, 1980).
Together, we applied these ideas to create a framework for understanding AI’s deeper implications for design, and in particular the education of architects.
Redesign Option Studio
Called “Redesign”, the 3rd year architecture course was concieved as a three-project journey:
- The first project was about demystifying AI
- The second project, was exploring the use of AI as a “rubber duck” to come up project idea, brief and requirements
- The final project, was exploring AI as a practical tool to resolve the detailing, materialization, and storytelling of the idea.
At each step, learning was discovered and constructed. A key departure in our pedagogy involved providing scaffolds that allowed students to define their own project briefs with complete freedom. Traditionally, architecture instructors — particularly at the undergraduate level — provide students with a set brief and fixed requirements. In contrast, we inverted this model, encouraging students to shape their own project based on personal interests in a “wicked problem” and architectural typologies.
A key departure in our pedagogy involved providing scaffolds that allowed students to define their own project briefs with complete freedom.
By semester’s end, the students’ understanding of AI had been transformed; When combined with their own intellectual foundations and architectural training, it became a living medium — one that invited critical thinking, ethical reflection, and promoted self-learning of technical skills. In this sense, AI was no longer something imposed upon learners; it was something they co-constructed, shaped to serve their evolving goals and deepen their understanding.
Project 1 — Immersion — Demystifying AI’s Grammar
The studio began with looking at a range of Generative AI tools such as GANs and diffusion models. Students explored what these tools do and why they matter to designers. We hosted a five-week workshop introducing large language models (LLMs) and basic machine learning concepts. We stripped away the jargon, revealing the underlying building blocks students could question and reshape.
By creating “Concept of Operations” (CONOPs) diagrams and “SWOT” analyses, they learned to see AI as a system integrated with its user. Students also were asked to consolidate their findings into presentations where they would consider how designers use different types of AI.
One student, for instance, used LLMs to “screenwrite” architectural ideas — imagining buildings as narratives and occupants as characters. This approach exemplifies a core constructionist principle: learning about AI should help us think differently about our own fields. Instead of passively absorbing information, students were actively constructing new mental models, discovering unforeseen design possibilities, and reimagining what architecture could be.
Project 2: Generative Briefing — AI as a Co-Creator
In the second project, students selected architectural archetypes — atriums, vaults, façades — and reimagined them to address contemporary challenges like climate change or mental health crisis. This stage embodied the “debugging” spirit of constructivist learning. Students wrote prompts, refined them, experimented with variations, and observed how AI outputs changed. They documented their iterative processes into concept exploration matrices and distilled their idea into the creation of “Hero Images.”
This stage embodied the “debugging” spirit of constructivist learning. Students wrote prompts, refined them, experimented with variations, and observed how AI outputs changed.
In one standout project, a student re-envisioned courtyard architecture to support cognitive well-being. By generating LLM prompts and systematically evaluating the results, they transformed the courtyard into different concepts for mental restoration, community engagement, and sustainability. Here, AI wasn’t producing tidy solutions — it was engaging in a conversation. The student continually asked, “How might we improve this idea?” AI became a partner in critical and creative thought rather than a mere tool.
Phase 3: Concept-to-3D — Materializing Ideas
The final phase shifted from the conceptual to the tangible. Students took their AI-informed designs and crafted physical prototypes — foam models, plaster casts, or 3D-printed structures. They adjusted parameters, worked within real-world constraints, and created narrative videos that gave meaning to their forms. As they manipulated materials, they also developed new ways of thinking, Since making and learning are intertwined.
Some groups found unexpected structural stability in vaulted forms without compromising elegance, while others devised modular building systems ready to adapt to evolving demands of society. Throughout this process, AI remained neither neutral nor perfect. It acted as a prompt, a collaborator, and a challenger, pushing students to refine their thinking and broaden their horizons.
Nuturing Reflective Practioners
By the semester’s end, it was clear the goal of teaching AI is not to produce technical users. Instead, it is about nurturing reflective practitioners who question their tools, critique their outputs, and shape technology to serve human values.
By the semester’s end, it was clear the goal of teaching AI is not to produce technical users. Instead, it is about nurturing reflective practitioners who question their tools, critique their outputs, and shape technology to serve human values.
We saw that iteration drives learning, that open-ended questions spark deeper inquiry, and that collaboration amplifies understanding. These findings align with constructionism’s core tenets: knowledge emerges through action, reflection, and social interaction.
Challenges remain. Balancing the time needed for developing technical fluency, while evolving foundational skills is no small task. Still, in future iterations of this studio, we will continue to refine our approach. After all, teachers do more than convey information — they create the conditions for invention.
Final Thought
In the age of AI, teachers must help students imagine futures not yet charted and supply them with the intellectual tools to construct those futures themselves. AI may be disruptive, but it is also ripe for play, experimentation, and invention.
In the age of AI, teachers must help students imagine futures not yet charted and supply them with the intellectual tools to construct those futures themselves.
By treating AI as an intellectual partner rather than a static instrument, educators can encourage students to understand and shape its role in design. We’re firm belivers that the role of the teacher is to create the conditions for invention. With AI in that context, we hope to empower young designers to explore, to critique, and ultimately to build a better future — one they will create on their own terms.
References
- Papert, S. (1980). Mindstorms: Children, Computers, and Powerful Ideas. Basic Book
- Ratto, M. (2011). Critical Making: Conceptual and Material Studies in Technology and Social Life. The Information Society, 27(4), 252–260.
Special Thanks
Everyone who participated in reviews and lectures:
- Mirko Daneluzzo
- Raffi Tchaikarian
- Aditiya Barve
- Christina Karvey
- Brian Muthaliff
- Hani Asfour
- Andrea Macruz
- Maria Nikolova
- Jianxi Luo
- Shuo Jiang
- Maya Przybylski
- Firas Safieddine
- Ava Mozafari
Digital Blue Foam
- Camiel Weijenberg
- Cesar Cheng
The team from Perkins and Will
- Daniel McTavish
- Andrew Frontini
- Nick Cameron
- Will Daravong
