DesignIntelligence talked with Phil Bernstein, lecturer in professional practice at the Yale School of Architecture, about the intersection of technology and professional practice; machine learning, artificial intelligence and big data; and how technology will impact the design practice in the future.
DesignIntelligence: Let’s talk about the intersection of technology and professional practice.
Phil Bernstein: Most of the work I’ve been doing lately is considering how technology is a catalyst or an enabler in changing the architect’s role in the system of delivery. I’m particularly interested in questions of value proposition, like how does technology change the architect’s value proposition in the overall economic system of project delivery? There’s a big disconnect between the value that we as architects actually bring to the building industry process and how we plug in and convert that value.
DI: How should leaders of professional service firms be thinking about technology?
PB: There are a few dimensions to this. The first is that as architects, we generate a tremendous amount of information that is broadly useful across the delivery spectrum. But we tend to hoard it and not see it as a value opportunity or as a way to plug into the system.
Second, “Big Data” is going to be a very necessary piece of design practice in the future. Data comes from a number of sources, like internal archives of digital information about projects, information that can be gathered during the course of the construction or operation of projects, large scale externally accessible databases, drone scanning and more. As architects, we must learn to understand and manage these large swaths of information.
Third, architects must think about and get ahead of the curve on machine learning and artificial intelligence, because it’s going to be an opportunity in the medium term. But in the longer term, it’s going to be a threat, in my opinion. There are many things that architects do that will be automated in the future. We need to have a stance on what that means and a stake in the process of developing and integrating it into the way we work.
DI: When you talk about a swath of what architects do today being eventually automated, can you give any examples?
PB: Let’s look at two or three potential categories of where machine learning-based automation might consume a part of the architect’s responsibilities. The first is in technical analytics, like code evaluation or cost estimating or energy analysis, which will rapidly transform from analysis formulas and algorithms that are managed by architects to machine learning algorithms that tell us the answers.
The second category is around information management—i.e., orchestrating and managing the flow of data and information among the design team and the construction team during a project that is largely the architect’s responsibility as part of coordinating the work.
The third category is to look at the entire spectrum of traditional basic services and estimate what percentage of those fees are just straight up and down production work, like producing standard documents, schedules, details, etc.; production items that are relatively straightforward to automate. We’re already seeing the early automation of that in BIM—does anyone create a door schedule by hand anymore?
The next wave is BIM meets machine learning, and so we’re likely to need fewer people to do all of that production work. What does that mean?
In an interesting book called Only Humans Need Apply, Thomas Davenport and Julia Kirby argue that the professions will be besieged by artificial intelligence. They recommend that we get out in front of the problem and start designing the parts of the practice that we think are best supported by machine learning and help chart the relationship between the humans and the machines.
DI: Can technology help us create new value propositions?
PB: Yes. On many different dimensions. The two I think are most interesting and provocative is first, the value of design information and its utility in the overall delivery chain. The second is this question of analysis and prediction. Algorithms are going to make the world a more predictable place within a certain set of limited, highly rational constraints. If that’s the case, how do architects leverage that predictability to increase their value proposition? In other words, instead of just buying my services as a low-price commodity, if I can make a commitment that X is going to happen, then I’d like to get paid when X happens.
DI: Do you believe there’s a core of what architects contribute that can never be replaced by a machine?
PB: Yes, primarily in the ineffable aspects of good design. But I wonder whether the normative pressures of the delivery system may overwhelm the system. There’s a synthetic, intuitive, innately human quality to solving complicated problems that I don’t think machines will ever replace. But the opposite side of that coin is there’s a certain number of highly normative buildings where nobody’s looking to solve a particularly complicated problem. They’re just trying to get between A and B as quickly as possible. Soon it won’t be hard to design a strip mall with a machine learning algorithm.
DI: With all of these technology issues and opportunities that will shape professional practice in the future, can the architecture profession alone solve these problems? Or will it require participation of others in the design and delivery process?
PB: Yes, by definition, anything that is transformational or even interesting that happens in the building industry is a multi-disciplinary problem. But when there’s a complex, ambiguous, multi-dimensional problem that requires a synthetic answer, architects are usually the ones who step in to that environment and help formulate the answer. That’s what we’re trained to do and where we could take a leadership role in moving things forward.
DI: Does cyber security become a part of the overall public health, safety and welfare mandate?
PB: I think so, but not so much from the perspective of data protection. What we’re going to find over the next several decades is that buildings, cities, cars and our phones are enormous data collectors that create big sources of information from which insight can be derived. What do we do with this insight? What is its implication on designers and how we build and how we live in the buildings that we build?
As a profession, we need to get in front of all of this, because data and technology will be an important and necessary piece of design practice in the future.