AImoji. Process Studio (Martin Grödl, Moritz Resl) created these AI-generated emojis for the Vienna Biennale for Change 2019, using a Deep Convolutional Generative Adversarial Network (DCGAN). According to the studio, “with each AImoji, new, hitherto unknown ‘artificial’ emotions come to life that challenge us to interpret and interact with them.” Courtesy Princeton Architectural Press.

The following is an adapted excerpt from Helen Armstrong’s new book, Big Data, Big Design: Why Designers Should Care About Artificial Intelligence.

Why should a designer care about machine learning (ML)? Fair question, right? ML consists of algorithms—in essence a set of task-oriented mathematical instructions—that use statistical models to analyze patterns in existing data and then make predictions based upon the results. They use data to compute likely outcomes. But what do these algorithms and predictions have to do with you? The answer grows more self-evident by the day.

Machine learning is everywhere and has already transformed the design profession. To be honest, it’s going to steamroll right over us unless we jump aboard and start pulling the levers and steering the train in a human, ethical, and intentional direction. Here’s another reason you should care: you can do amazing work by tapping the alien powers of nonhuman cognition. 

Machine learning has changed the way humans relate to machines by enabling them to communicate with tech via language, gesture, movement, emotion, etc. These same capabilities will enable designers to engage with creative tools in more intuitive ways, supplanting the mouse, trackpad, and touchscreen. Simply asking software to perform an action—rather than clicking and dragging through a menu to find the right tool—will, for example, allow designers to bypass hours of busywork, not to mention perusing dense tutorials. Perhaps the very concept of a “tool” will grow irrelevant. The more natural and personalized the interaction, the more creative software might feel like an extension of ourselves—and our individual creative approaches—rather than a separate clunky software package.

Escape the Cubicle 

If design tools combine relational interaction with artificial intelligence’s (AI) growing awareness of context, designers could, at long last, escape desk and screen. We could design for the world as we stand in the world, creating while situated in augmented physical space. Silka Sietsma, Head of Emerging Design at Adobe, asserts, “We’re at the event horizon of a new era of spatial computing—a world where digital experiences mesh with physical reality. Immersive, 3D technologies like AR (augmented reality) and VR (virtual reality), along with voice and embedded sensors, are all converging into a new medium, powered by artificial intelligence.” It is within this new medium—this confluence of physical and digital, that our future design practice will evolve.

How else will ML impact design practice? Patrick Hebron, author of Machine Learning for Designers, suggests that we consider the future in terms of “scaffolding complexity.” While reflecting on CAD systems and the future of creative tools, he points out: “These tools make it possible to conceive of systems that are too grand and complex for any one individual to keep all of their big picture goals and specific details in mind simultaneously.” Hebron also notes, “The Florence Cathedral took about one hundred and forty years to go from initial conception to project completion. A much more complicated and recent building, the Burj Khalifa, took about five years.” Our tasks and aesthetic goals, he asserts, will continue to evolve as ML enables us to enter terrain that we could not even envision without AI. “Machine intelligence,” he explains, “will enable creatives to do even more and to think even bigger.”

Return of the Centaur

In such a vision, humans and intelligent machines work together to arrive at solutions unattainable by either alone. We can refer to this as a centaur: part human, part machine intelligence, one entity.  Rather than automating away the designer, designers join forces with AI, augmenting their abilities with ML—a fusing of intelligences. Matt Jones of Google AI argues that to truly take advantage of AI, we must accept its alien nature.

Alternative models of “intelligence” exist already in the natural world—specialized forms of cognition distributed across organisms, nerve cells, and root-fungi networks rather centralized in a single human brain. Myriad recent books have raised popular awareness of these alternatives, such as Peter Godfrey-Smith’s book, Other Minds: The Octopus, the Sea, and the Deep Origins of Consciousness and Peter Wohlleben’s treatise, The Hidden Life of Trees. As posited by posthuman theorists like N. Katherine Hayles and Donna Haraway, we should expand our understanding to other forms of cognition as we coevolve with our tools. In essence, we must recognize that integrating ML into design practice will not feel like adding a supersmart fake human to our creative team but, instead, will be something else entirely. Like bacteria, trees, or earthworms, AI will think differently than we do. 

The Combo, Please

Since the 1960s, we have imagined that AI will take over form-making, serving up a multitude of form variations from which a designer can simply choose—a fast forwarding of the design process. But, it turns out, the most powerful application of ML is not speeding up our process to arrive at the same kind of conclusions. The most powerful applications combine machine intelligence with human intelligence to take us along new paths entirely. 

Artificial intelligence researcher Janelle Shane puts it simply: “Working with AI is a lot less like working with another human and a lot more like working with some kind of weird force of nature.” This force of nature will tirelessly work toward exactly the goal that we give it, so we have to figure out the right goal. And we shouldn’t expect it—or want it—to solve the problem like a human. Shane points to a project by David Ha, a researcher at Google Brain, in which Ha asked an AI to assemble some parts into a robot to move from Point A to Point B. Rather than solving the problem by assembling a nimble robot, as Ha intended, the AI combined the parts into a tower that could just fall over and land on Point B. As Hebron comments, “The world is full of human thinkers. And if we want human thinking, we should probably go to humans for it. There are a lot of them.” If we don’t waste time trying to force AI to think like a human, we can arrive at Point B—and Points C, D, and E—in fresh, alien ways. 

“Working with AI is a lot less like working with another human and a lot more like working with some kind of weird force of nature.”

Novel AI strategies, however, mean little without perspective and purpose. Humans do need to be part of the equation. Remember the human-AI chess teams that triumphed over solo humans or solo AI competitors? The confluence of human and machine is key. As Shane explains, “The AI has no understanding of the consequences.” Humans bring that understanding to the equation. We human designers must be there to frame the right problems—the problems that will move us toward future points that truly benefit humanity.

The Future?

The future is…fraught. Our profession stands on the cusp. Designers must strive to understand ML capabilities, so that we can engage with it as a design material and a creative force. If we do not, we will fall victim to it. We will create within the parameters that the technology sets for us, rather than the other way around.

Through ML we have amazing potential to provide emotional insight to those on the autism spectrum, to reduce gender and racial bias in hiring and lending practices, to springboard creatives into unexpected, wicked problem-solving spaces. However, we can also do the exact opposite—exploit the vulnerable, bias the future by relying upon the past, replace humans by automating away skills that we want and need to maintain autonomy and agency, relegate essential choices to a technology that has no understanding of human consequence. 

Questions around AI and humanity have been hotly debated since at least the middle of the twentieth century. As design professor and historian Molly Wright Steenson points out, “If we understand that we’ve been asking these questions for a long time, we might have better expectations about how hard it is to find answers.” These are complex questions with wide implications. 

The terrain is tricky. The future is uncertain. Exciting? Yes. Terrifying? Yes. We have many critical choices ahead. Let’s take on those choices together, thoughtfully, one design at a time.

Read more in Helens new book, Big Data, Big Design: Why Designers Should Care About Artificial Intelligence, now available from Princeton Architectural Press.