e-SPAN Newsletter v032: AI Pedagogy & Research

Images painted by artificial intelligence including a field, balloons, a scooter and a woman.

Different image generation techniques, Manuel Rodríguez Ladrón de Guevara (PhD-CD '23), thesis research.

Dear School of Architecture Community,

This month I’m excited to share with you some of our faculty’s and students’ recent research and pedagogy—as well as some of my own thinking—around artificial intelligence (AI).

As I’m sure you’re aware, AI is evolving daily; to some extent, any prediction about its future is foolhardy. But AI was pioneered at CMU, and we take seriously the responsibility and opportunity to explore and shape how this evolving technology impacts architecture and the built environment. To that end, we have made artificial intelligence one of our three priority pedagogical challenges (along with social justice and climate change) addressed by all of our degree programs.

My own position on the advent of greater intelligence in the functioning of the built environment is cautiously positive. I see great potential, for example, in harnessing AI in the built environment’s response to climate change. Much of this intelligence will come from the Internet of Things (IoT), as AI-assisted sensors and actuators embedded in our architecture provide real-time data on its performance and choreograph its response. This will generate profound benefits in energy consumption, safety and environmental health.

Likewise, some promising work in machine learning seeks to impart to robots some of the embodied knowledge acquired by craftspeople through repeated engagement with materials—thus both extending the reach and pushing the boundaries of venerable craft techniques. And on construction sites, co-robotics and assistive technologies—like self-driving vehicles—carry the potential to increase construction safety and efficiency.

Like many of you, though, I also harbor deep concerns about a host of challenges inherent in AI.

The general public has recently been introduced to large language models (LLMs) like ChatGPT, Midjourney, DALL-E and Sora—which generate text, images or video from text prompts. Trained on vast datasets of images and texts, these tools provide users with a means to develop new content.

Many Carnegie Mellon Architecture alumni and friends have expressed both curiosity and dread about the ways these tools are redefining architecture education. I’m pleased today to share with you a small sample of the creative and exploratory ways our students and faculty are deploying them in the design process.

For educators, LLMs pose (at least) two important ethical dilemmas. First is the problem of authorship: is the work I’m reading or seeing—essay or image—a student’s intellectual production or the output of an LLM? And second is the concern for uncompensated content: if I use an LLM in my class or studio, am I freely contributing students’ intellectual production to further training that LLM?

I’m grateful to work alongside designers and educators who are moving through and beyond these dilemmas, using AI to analyze and solve a variety of design problems. As you’ll see below, our faculty and student work ranges from ideation and rendering of architecture in studio projects, to researching innovative approaches to fabrication and craft, and also engages the question of whether AI itself can be creative.

We will continue to share more in the future, and we welcome your questions and observations on the place of AI in architecture and design. As ever, we are committed to helping our students lead: in developing new technologies, and in adopting them ethically, thoughtfully and creatively.

Over the next year we'll continue to explore AI through our public programs and pedagogical discussions. In the fall, our AIAS student organization will be hosting the AIAS Northeast Quad Conference, Des[ai]gn: AI in Architecture, which will share and expand our and CMU’s cutting edge research in this area. Other conferences and events are being planned across the College as well.

As March turns to April, we’re looking forward to Spring Carnival, taking place April 11-14. Carnegie Mellon Architecture will host an alumni get-together at the Spring Carnival Pavilion Dedication & Alumni Open House on Saturday, April 13, 11:30am-1pm in front of the Carnival Entry Pavilion—designed again by our NOMAS chapter. Over the last two years, NOMAS students have masterfully represented their values through pavilion design, and this year’s is just as compelling. I look forward to seeing you there.
 

Omar Khan
Head of School


AI for online engagement in interactive fabrication


 

Two renderings of a cityscape generated by artistic artificial intelligence.

Side-by-side: artificially generated paintings, Manuel Rodríguez Ladrón de Guevara (PhD-CD '23), thesis research.


AI for freeing arduous and mechanical human labor, and by extension— potentially making humans more creative

Manuel Rodríguez Ladrón de Guevara (PhD-CD '23)

In my PhD work (in very simple terms), AI tries to generate novel artistic painting styles just as humans do—that is, having a perception of a scene, transforming that perception following some internal mechanism, and generating a painting version of the scene. In another chapter of my thesis, AI uses a similar principle to depict an avatar version of a person's face, maintaining the identity of the person.

The field of generative AI and computer graphics is broad and complex, and requires a lot of time and resources. The goal of AI systems is to generalize their results to unseen cases. Measuring how well an algorithm generalizes the work is a good way to understand pitfalls or success.

In my case, I became very interested in the potential that AI has in design and design languages. There are many questions remaining, such as whether AI will actually be creative (the answer tends to be negative, but it also depends on a definition of "creativity," as discussed in my thesis). 


 

Artistic rendering of structure from above.

Mobile Home Hybrid, Ella Moon (B.Arch '26), from "Mobile Home" studio with Jared Abraham at CMU Architecture.


AI for the rapid production of architectural studies, testing formal variations, and expanding design ideation

Jared Abraham shares student work from his Mobile Home 2nd Year Options Studio

The 2nd Year Options Studio Mobile Home utilized a hybrid design approach, combining traditional analog media with AI image-generation tools. Students first produced hand drawings and physical study models. Images of students' drawings and models were then combined with text descriptions in Midjourney (an AI image generator), producing an array of digital images representing formal variations based on the analog inputs. Students were then forced to grapple with the question of what (if anything) should be incorporated into their designs.

Work from Benni Guo (B.Arch '26), Ella Moon (B.Arch '26) and Henry Youngren (B.Arch '26) is shared here.


 

The results of six prompts related to the built environment in English, Amharic, and Amharic script, eight months apart entered in Midjourney.

The results of six prompts related to the built environment in English, Amharic and Amharic script, eight months apart entered in Midjourney. Image by Vernelle A. A. Noel.


AI for a visual argument

Vernelle A. A. Noel, Lucian and Rita Caste Assistant Professor in Architecture

In one instance, I used machine learning to imagine new design possibilities for dancing sculptures and traditional characters in the cultural design event of Carnival in Trinidad & Tobago. I then designed and built virtual immersive spaces for anyone to visit to see the work and learn about the Carnival and machine learning. 

In another instance (represented by above image), I raised questions about semiotics, semantics, syntax and society in text-to-image generator tools by entering six prompts related to the built environment in six different languages, eight months apart. The work is a provocation to consider the implications and perturbations of understanding the world through these tools.

Among my current questions: How is AI reconfiguring our cognition as designers and animate beings in the built environment? And what does it mean for architecture when we privilege technological knowledge rather than social, historical and cultural learning?

Remote video URL

Spring Events

We look forward to welcoming you to the following upcoming events during Spring Carnival weekend in April.

Spring Carnival Pavilion Dedication & Alumni Open House | Saturday, April 13, 2024 | 11:30am-1:00pm
The School of Architecture invites alumni, students, faculty, staff and friends to join us for the Spring Carnival Pavilion Dedication & Alumni Open House on Saturday, April 13, 2024 from 11:30am-1:00pm. Drop by the official CMU Spring Carnival pavilion to meet School of Architecture Head Omar Khan and current students who designed and built this NOMAS-led project.

Retirement Reception and Roast in Honor of Steve Lee | Saturday, April 13, 2024 | 5:30-7:30pm
You're invited to a Retirement Reception and Roast in Honor of Steve Lee during Spring Carnival weekend on Saturday, April 13, 2024 from 5:30-7:30pm. The reception is at 5:30pm, roast begins at 6:30pm. Questions? Reach out to Kristen Frambes. RSVP here.


Alumni News & Updates

We invite all Carnegie Mellon Architecture alumni to keep us up to date on their awards, professional milestones and more. Send us your updates with a brief description and link to more information.

Share your news

Susie Kim (B.Arch ‘23) has been selected as the winner of Carnegie Mellon Architecture's 2024 Delbert Highlands Travel Fellowship for her proposal “Hiroshima as Tabular Rasa: a new modern urban ethos in response to crisis and uncertainty.” The fellowship supports alumni in the study of collections belonging to locales to promote the professional development of awardees and contribute to the richness of our surroundings. The next call for applications is in spring 2025.

Zain Islam-Hashmi (B.Arch '19) was named to Carnegie Mellon University’s 2024 Tartans on the Rise. Now in its third year, Tartans on the Rise celebrates alumni who graduated in the last 10 years and are making an impact on their organizations and in their communities, across the nation and around the world through leadership, innovation and career accomplishments.

Jason Shannon (B.Arch '03) and Paola Yañez (B.Arch '03) and their practice J_spy have developed a project to address the affordable housing crisis. J_spy decided to go an untraditional route and offer a free house design, Free Haus. Anyone can download the architectural drawing set and use it to build their own house. More details, and the plans, can be found on their website.

Raymond L. Gindroz, FAIA, alum and former faculty member, published Taking a Pen For a Walk. In this book, he shares over two decades of experiences of place through drawing. Accompanied by concise yet charmingly anecdotal and informative commentary, this extraordinary collection of over 500 travel sketches invites us to discover the hidden language that buildings speak, one deeply rooted in both history and the human experience.