Alpine Villa. Image © SPAN (Matias del Campo, Sandra Manninger)

What is artificial intelligence? How can we maximize the benefits of AI while minimizing the risks? Is the role of the architect safe from AI?

In any conversation about the impact of artificial intelligence on the architectural profession, such questions are at the forefront of many human minds. They are also a sample of the series of questions ChatGPT posed to the architect and educator Matias del Campo. In a recent experiment for Archinect, the University of Michigan associate professor and SPAN co-founder switched roles with the popular chatbot to share insights on his background in artificial intelligence and his thoughts on its future application to the design process. The full conversation is republished below.

This article is part of the Archinect In-Depth: Artificial Intelligence series.

ChatGPT: I would love to hear more about you and your background. Would you be willing to share some details about yourself, such as your education, professional experience, and any other relevant information that you feel comfortable sharing? This can help me better understand your perspective and expertise and provide more context for our conversation.

Matias del Campo: As a registered architect, designer, author, and educator, I have had the privilege of holding various roles in the field of architecture. Born in Chile and raised in Vienna, Austria, my multicultural background has greatly influenced my perspective on architecture and design. Growing up in Vienna, I was exposed to a rich history of art, architecture, and design, which has left an indelible dent in my work. 

I am currently an Associate Professor at Taubman College for Architecture and Urban Planning at the University of Michigan, and the director of AR2IL: The Architecture and Artificial Intelligence Laboratory at UoM. My research interests are focused on advanced design methods in architecture, leveraging AI techniques through collaborations with the Computer Science Department, Data Science Department, and Michigan Robotics. While I am grateful for the opportunity to contribute to this exciting field, I am also keenly aware of the limitations of my knowledge and expertise. Thus, my tendency to seek interdisciplinary collaboration. 

This hot friction between historic forces and the resistance to it is what fuels most of the projects I do.

My education at the University of Applied Arts in Vienna was a formative experience that shaped my approach to architecture. The university's commitment to innovation and experimentation encouraged me to explore new ideas and push the boundaries of what was possible. I felt, and still feel, a sense of powerful suspension between the gravitational forces of the Viennese tradition (Adolf Loos, Otto Wagner, et al) and the forces pulling me to resist that tradition (Hans Hollein, Coop Himmelb(l)au, and others). This hot friction between historic forces and the resistance to them is what fuels most of the projects I do.  

As a co-founder of SPAN, I am proud to work alongside a talented team of professionals who share my passion for innovation and the critical interrogation of disciplinary questions. Together, we strive to deliver projects that meet the needs of our clients and contribute to architecture as part of both the built environment and as a force shaping human culture. In my work, I approach architecture with a deep respect for the knowledge and expertise of my colleagues. I am honored to be part of this dynamic field and I am committed to continuing my own growth and development as a professional.

I would love to know more about your interest in the intersection of architecture and AI. Could you elaborate on what specifically caught your attention about this field and why you find it compelling?

In 1997, my partner, Sandra Manninger, introduced me to Prof. Robert Trappl, the director of the OFAI, the Austrian Institute of Artificial Intelligence. We had several interesting conversations on the use of AI in general and its potential for architecture design. In 2006, Sandra and I gave our first workshop on machine learning at the Angewandte in Vienna, in collaboration with Dr. Arthur Flexer of the OFAI. Around 2016/ 2017, technology finally crossed the threshold where more extensive research could begin. 

The integration of AI in architecture has the potential to create a more collaborative and inclusive design process.

The introduction of generative adversarial networks (GAN) made exciting design research possible. I was always attracted to the idea of exploring whether sensibility and agency can emerge from a non-human environment, as described in my book Neural Architecture. Not to replace humans as actors but to expand the creative palette. This being said, my interest in the potential of AI in architecture was piqued by the prospect of exploring new design possibilities that were previously unavailable to us.  

In the early stages of our exploration, Sandra Manninger and I were fascinated by the idea of developing a nonhuman design partner that could assist us in generating designs that were both innovative and practical. Our initial workshops on machine learning at the Angewandte provided an opportunity to collaborate with experts in the field of AI and to gain a deeper understanding of the potential applications of these technologies in architecture. 

As technology continued to evolve, we were able to leverage more advanced artificial intelligence techniques in our research. The introduction of GANs opened up new possibilities to explore the creative potential of AI in architecture. We were able to generate a wide range of architectural designs, from simple structures to complex geometries that would have been impossible to create through traditional design methods. The findings of our research were also published in my book, Neural Architecture. The book explores the possibilities of AI in architecture and its potential to create a more responsive, adaptive, and sustainable built environment, whilst positioning these aspects in a specific cultural framework. 

I argue that the integration of AI in architecture has the potential to create a more collaborative and inclusive design process, where the architect works in tandem with an intelligent design partner to generate solutions that are informed by data and responsive to the needs of the occupants. As my colleague Mingyan Liu, the chair of computer science at UoM, pointed out in the ‘Data Justice, AI and Design’ colloquium: Instead of calling AI ‘Artificial Intelligence,’ we should call it ‘Assisted Intelligence.’ My experience with this tool supports this notion. It is a machine that expands the human capabilities of intellectual reasoning and creative interrogation of a problem. To that end, I would argue that working with artificial intelligence is the very first 21st-century design method in architecture. 

By using the predictive abilities of AI, architects can potentially improve the living conditions of millions around the globe. 

In AR2IL, The Architecture and Artificial Intelligence Laboratory at the University of Michigan, the main trajectory of the research continues to focus on exploring the potential of AI in architecture. There are two particular areas in which artificial intelligence excels: optimization and prediction — both are areas with widespread application possibilities in architecture. For example: In collaboration with the Computer Science Department and Michigan Robotics, we are exploring the use of AI in areas such as building performance analysis, floorplan optimization, sustainable demolition methods, and autonomous 3D model generation. 

I believe that the integration of AI in architecture has the potential to transform the field, enabling us to create structures and spaces that are not only functional and aesthetically pleasing but also responsive to the needs of the occupants and the environment. By using the predictive abilities of AI, architects can potentially improve the living conditions of millions around the globe. As the field continues to evolve, I am excited to see where the future of AI in architecture will take us.

Can you provide your understanding of what AI means? How do you characterize artificial intelligence and its various functionalities?

AI is an umbrella term used to describe several approaches. In general, we speak of artificial intelligence when machines mimic cognitive functions that humans associate with other human minds, such as learning and problem-solving. AI is a complex field that encompasses various approaches and techniques. At its core, AI is concerned with creating intelligent agents that are capable of perceiving their environment and taking actions that maximize their chances of achieving their goals. This definition, which is widely accepted in the field of computer science, highlights the central role that agents play in AI. In other terms, AI can take many different forms, ranging from rule-based systems that use logical inference to make decisions to machine learning algorithms that can learn from data and adapt to changing circumstances. There are multiple AI systems that are also capable of natural language processing, which allows them to understand and generate human language. 

The ultimate goal of AI is to create machines that can exhibit intelligent behavior that is comparable to that of humans. This involves developing algorithms and techniques that can mimic cognitive functions such as learning, problem-solving, perception, reasoning, and decision-making. According to Pedro Domingos, there are five different tribes within AI research: Symbolists, Bayesians, Connectionists, Evolutionaries, and Analogizers. The most successful ones are currently the connectionists with the idea of the neural network (NN). NNs allow us to mimic processes in our brain, the behavior of our biological neural pathways. Insights from neuroscience are cast into math, which allows us to do algorithms using our mind as a blueprint, which we can then turn into machine learning processes. 

It is important to note that artificial intelligence is not a replacement for human creativity and expertise.

Of course, we can only imitate what we understand, which defines the limits of AI these days. We have a good grasp of how dreaming or hallucinating works; thus, we can use these aspects in neural networks. So, yes, machines can hallucinate, but they will not develop consciousness any time soon because we do not know how consciousness works in its entirety.

In the context of architecture, AI can be used to automate various tasks such as generating design options, evaluating environmental performance, and optimizing building systems. By leveraging AI, architects can explore more design possibilities and make more informed decisions. However, it is important to note that artificial intelligence is not a replacement for human creativity and expertise. Rather, it is a tool that can assist architects in their work and enhance their ability to create innovative, provocative, and sustainable designs.

Given the rapid pace of AI development and the potential impact it could have on our daily lives, I am curious to know whether you generally feel more optimistic or pessimistic about its future. Are there any specific factors or experiences that have shaped your perspective, and how do you think we can maximize the benefits of AI while minimizing the risks?

I am somewhat agnostic about this question. I agree with Manuel Delanda, who describes the abilities of objects as 'potentialities.' Ultimately, every tool can be used as a weapon. For example, a knife can be used to kill someone or cook a delicious meal. Similarly, artificial intelligence can be used with malicious intent or to aid humanity. This is not up to AI but to the person programming it. If at all, it is the human operator that infuses biases, prejudices, and evil intent into the neural network. We have to pay attention to how artificial neural networks and the respective datasets are created. I have repeatedly stated that if we as architects do not engage with AI, someone else will do it for us (investors, developers) — shutting us out from decisions on how to use AI in architecture.

We are far from understanding how the jumbled pile of neurons in our skulls can form consciousness or even a thought.

There has been ongoing debate and speculation about the possibility of artificial intelligence achieving consciousness or self-awareness. I am interested to hear your thoughts on this topic. Do you believe that it is possible for AI to develop a subjective experience or consciousness similar to humans, and if so, under what conditions?

I would be cautious with this question. Just consider that science still needs to understand consciousness in all its details. There are interesting theories, such as Anil Seth’s ideas about consciousness, perception, and awareness. Still, we are far from understanding how the jumbled pile of neurons in our skulls can form consciousness or even a thought. That being said, I don’t think we know enough about the topic to be able to encapsulate consciousness in an algorithm. I guess we are close to making machines that can think, but that does not mean that the machine knows that it is thinking, which would be self-awareness.

Given the rapid advances in artificial intelligence technology, there is increasing speculation about the range of tasks and occupations that AI can perform. I am curious to hear your thoughts on what types of jobs you believe AI is already capable of doing, and which industries have been most affected by AI automation.

Here, it is necessary to differentiate between automation and AI applications. Automation, which means using (industrial) robots, is gaining traction in various fields, from logistics (warehouse robots) to medicine (operation robots) to hospitality (burger-flipping robots, etc). AI, and here I would be more precise and say learning systems based on neural networks (NN), are practically everywhere. Facial recognition is implemented in mobile phones, and NLP (Natural Language Processing) is implemented practically everywhere. NN algorithms suggest books to you via Amazon, and YouTube recommends videos based on a learning algorithm that analyzes your watching behavior. Adaptive maps guide an algorithm that drives your rideshare somewhere. Facebook filters your feed according to your behavior and shows you posts that increase your attention to keep you on the page. Traffic lights are increasingly controlled by neural networks. Medical doctors use NN’s in diagnosis; lawyers use them to find precedent cases faster, etc. 

There are many more examples. In short, AI is everywhere. The question is not about which jobs they can do, because they can do almost any job, but about how humans will adapt to it.

Looking toward the future, what new opportunities and advancements do you think AI will bring to the design and construction industry? How might AI enable new forms of creativity and innovation in the field, and what skills and knowledge will be necessary for design and construction professionals to thrive in an AI-driven environment?

AI applications in architecture and design are emerging. The tamed problem works in areas of spatial and structural optimization using neural networks. The wicked problem works on things like sensibility, culture, ethics, and aesthetics of design tasks. This is what I call a 'Neural Architecture.' Machine learning applications are getting increasingly embedded into off-the-shelf software, there are already examples out there. We will see the same trend in architecture and engineering software packages, to the point that users might not even be aware that they are actually using an AI in their work.

Machine learning applications are getting increasingly embedded into off-the-shelf software

Can you give us an example of a project where AI was used in design and construction?

The Robot Garden, Ford Robotics Building, University of Michigan, Ann Arbor, USA, 2019

How would you feel about having an AI colleague?

I already have not one but several AI colleagues, and I feel very comfortable with them.

Are our jobs safe from AI? Why are architects skeptical about AI? Given the increasing prevalence of AI technology and its potential to automate various tasks and processes, there is growing concern about the future of employment and the impact of AI on various professions. I am interested to hear your thoughts on the topic of job security and AI, particularly within the architecture field.

Humans fear what they don't know, and this is very true for the entire context of AI. Perhaps the fear is fueled by our tendency to anthropomorphize things, projecting our own abilities and failures onto machines. Perhaps it is due to overexposure to dystopic science fiction movies like The Terminator or Ex Machina. However, when I teach AI to architecture students, job safety questions come up immediately (and understandably, it is their future after all). Once they understand the significant amount of work required to prepare data, code neural networks, train them, fine-tune algorithms, and ponder the ethical questions, they no longer ask what you are asking me here. 

Our ability to adapt has been vital to our species' survival, and we need to be prepared to adapt to AI's impact.

Like any new technology, AI may cause some jobs to disappear and transform. Other new jobs will emerge in the process. Our ability to adapt has been vital to our species' survival, and we need to be prepared to adapt to AI's impact. Additionally, I am fascinated by the potential effect AI could have on our economic systems, which currently are rooted in theories from the late 19th and early 20th centuries. It is time to rethink what we mean by 'work' and move beyond traditional notions of 'jobs.'

I am looking forward to the economic theories that AI will undoubtedly provoke, making life easier for everyone.

The emergence of artificial intelligence is considered a disruptive technology that can potentially interfere with the traditional roles and responsibilities of architects. In light of this, it is important to consider to what extent AI will affect the field and whether it is necessary to redefine the proper function of both AI and humans. With the increasing use of AI in architecture, it is important to evaluate how AI can be integrated into the field without replacing or diminishing the value of human expertise. Furthermore, it may be necessary to redefine the roles and responsibilities of architects in the age of AI and establish a clear understanding of how humans and AI can work together to create better designs and more efficient processes. What are your thoughts on this? Should the use of AI be limited?

AI technology can be viewed as an extension of the architect's mind and abilities, providing a new tool for designers to create innovative and efficient designs. It is important to consider who should be responsible for regulating the use of AI technology in architecture. Who should decide that? The government? The architects' chamber? Although regulations may be put in place to ensure safety and prevent misuse, they can also be counterproductive to progress in design, science, and engineering. 

Instead of focusing on regulations, it is crucial to explore how AI technology can be utilized to benefit the field of architecture and enhance human creativity. It is also important to acknowledge that humans may not be the only creative force in the universe, and that AI technology may have the potential to offer new perspectives and solutions that were previously unattainable. However, it is still too early to define the role of AI and humans in architecture, and there may even be a possibility that our current understanding of AI is flawed.

It is also important to acknowledge that humans may not be the only creative force in the universe.

Our understanding of intelligence and the processes that underlie thought, intuition, and memory is still limited, which means that our ability to create truly artificial intelligence may also be limited. As such, it is important to continue exploring the possibilities of AI technology while remaining humble in our understanding of what it can achieve.

How will AI shape the future of architecture?

AI is already beginning to shape architecture by enabling architects to design more efficiently and effectively. With AI, architects can automate tasks such as generating floor plans, optimizing building performance, and identifying design flaws. At the same time, it allows architects to mine the vast repositories of architectural knowledge that our discipline has collected over millennia. Meaning that we can be archeologists of the future, connecting the many dots (data points) created by our discipline to find solutions to contemporary problems. These problems include pragmatic aspects as much as the idea of architecture as a cultural artifact — evidence of architecture as a combination of material and symbolic culture. 

The use of AI in architecture has the potential to revolutionize the way buildings are designed, built, and used

In addition, AI can assist with data analysis, allowing architects to make more informed decisions based on data-driven insights. This is where the area of data science can be of enormous help to architecture. For example, AI can help architects analyze building performance data to determine which design features are most effective in reducing energy consumption. 

Overall, the use of AI in architecture has the potential to revolutionize the way buildings are designed, built, and used, leading to more efficient and sustainable structures that better serve the needs of the people who inhabit them and that can truly express the culture of our contemporary age.