Tackling a Biden Challenge with Artificial (and Human) Intelligence

February 10, 2021

Managing Partner, NBBJ

Editor’s Note: This post initially appeared on Architect’s Newspaper. It was co-authored by NBBJ Managing Partner Steve McConnell, Yale University Professor Phil Bernstein, journalist Cliff Pearson and senior AI researcher Dr. Mark Greaves, Ph. D.

 

Tucked within President Biden’s year-one legislative agenda on climate change is a call to build “zero net energy buildings at zero net cost.” This is a bold challenge that resonates powerfully in both the architectural profession and America as a whole. Like many great challenges, it will require a transformation in the way a broad range of disciplines work to shape the built environment.

The benefits of meeting Biden’s challenge are huge. According to the nonprofit organization Architecture 2030, “The urban built environment is responsible for 75% of annual GHG [greenhouse gas] global emissions: buildings alone account for 39%. Eliminating these emissions is the key to addressing climate change and meeting the Paris Climate Agreement targets.” So, the ability to cost-effectively produce zero net energy buildings would over time make a massive positive impact on our climate problems.

The root of the challenge’s difficulty is that designing and building great buildings is already a classic “wicked problem.” Wicked problems are defined by imprecise goals, incomplete knowledge, deeply interconnected subproblems, and the need to continuously make “best guess” tradeoffs. Instead of right or wrong answers, wicked problems require us to think in terms of better or worse solutions. Biden’s challenge adds substantially to the difficulty of these tradeoffs in architectural design, and further requires that we do this at zero net added cost.

Good architecture emerges from successfully balancing the interests of all stakeholders in a building project, while simultaneously optimizing innumerable decisions about structure, mechanics, economics, and aesthetics. Adding a zero net energy requirement will likely result in either increasing the cost of design and construction, or cutting back on space or amenities.

We think it is critical that the zero net energy buildings envisioned by President Biden also make positive contributions as works of architecture and valuable parts of the urban fabric. Otherwise, we could end up with super-insulated, faceless boxes that reduce our carbon footprint and are cheap to design, but undermine the vibrant character of our neighborhoods and towns. The Biden challenge sits at the intersection of some very big issues, from energy efficiency and environmental justice to advanced building materials and lively urban communities. It’s inspiring, but daunting, to confront.

Fortunately, the design profession is evolving, as society demands more from the people in charge of our buildings, neighborhoods, and cities. Architects now work in cross-disciplinary teams and handle an expanding spectrum of tradeoffs involving environmental factors, complex client needs, elaborate regulatory requirements, and constantly changing prices and availability of building materials. Architects also routinely balance less quantifiable factors such as the health of impacted communities, societal goals for the built environment, and justice in labor practices across the supply chain. To achieve this, they rely on a combination of deep design knowledge, extensive experience in how different designs will ultimately function, and powerful computational tools that can illustrate the impact of various tradeoffs. Meeting Biden’s challenge in a cost-neutral manner, though, is beyond the capability of current tools and practices.

We believe that new developments in artificial intelligence (AI) are the key to conquering Biden’s zero net energy challenge. Just as AI has revolutionized fields as disparate as drug discovery and self-driving cars, new AI-driven architectural tools can provide the support needed to cost-effectively design inspiring zero net energy buildings. Modern machine-learning algorithms don’t blindly follow a set of preprogrammed rules, but instead develop their capabilities by analyzing large sets of examples. With more and more data and feedback, they perform better and better. The latest AI language models such as GPT-3 are trained on billions of sentences from the web and can generate astonishingly fluent essays from a simple prompt.  AI software can produce well-rounded stories from just a few pieces of information, competent poems and pictures from a few prompt words, and even satisfying music from a few snippets of melody.

Could an AI tool produce compelling zero net energy building designs at zero net cost all by itself? No. Architects wrestle every day with wicked problems that are essential to creating compelling building designs, and many of the important design tradeoffs they make cannot be defined tightly enough to train an AI algorithm. However, AI promises to free them to focus more on what they can uniquely do: bring that hard-to-explain flair and creative spark to solving difficult design problems.

AI will make it possible for architects to cost-effectively address the enormous complexity inherent in the Biden challenge, by analyzing huge amounts of data to rapidly present options for design teams to consider and refine. This is essentially what Spotify does when it recommends music we might like. In architecture, AI can accelerate the design process by identifying subtle patterns that are likely to satisfy a set of design requirements. It can rapidly generate plausible zero net energy configurations, accounting for a broad range of factors and constraints. Finally, AI can work with advanced simulation technology to help architects assess the effectiveness of various design solutions to satisfy the diverse constituencies for a building project.

AI promises to be a disruptive technology for architects, but it is not a total solution. In the end, design requires understanding and evaluating a series of tradeoffs and picking the best ones. Designing great buildings that inspire their stakeholders is a task that people do better than any algorithm. The art of architecture requires a creative spirit behind it, even as designers apply increasingly sophisticated digital tools to tackle the wicked, fantastically difficult problems of delivering compelling, zero net energy buildings at zero net added cost.

President Biden, we in the architecture and computation fields accept your challenge and look forward to working with your administration to transform buildings in America.

______________________________________________________________________________

Steve McConnell is an architect and managing partner at the global design firm NBBJ.

Phillip Bernstein is associate dean and professor adjunct at the Yale School of Architecture.

Mark Greaves is a senior AI researcher.

Clifford Pearson is a journalist who covers architecture and urbanism.

The views expressed in this article are those of the authors and do not necessarily reflect the positions of their employers.

Image by Silvestri Matteo / Unsplash

Share this:  envelope facebook twitter googleplus tumblr linkedin
Follow nbbX