Now that robots look like they are in fact serious about taking some of our jobs, us humans can at least reassure ourselves that we are better at creative tasks like design. But is that, in fact true?
Consider the new iPhone 7. It is thinner, faster, has a better camera and water resistance, and (wait for it) comes in a new color — but arguably, Apple’s reductionist and predictable approach to device evolution, now means that an algorithm could potentially optimize the next iPhone as well, if not better, than a mythical designer like Jony Ive.
Machines excel at design challenges when the problem is essentially one of managing trade-offs. Known as computational design, algorithms can generate optimal solutions based on certain parameters. A very useful approach, it turns out, if you want to design something like a building.
Consider Seattle-based architecture firm, NBBJ, for example. They have designed workspaces for some of the world’s biggest technology brands — Google and Amazon in the US, and Alipay and Tencent in China. Unlike a traditional design firm, NBBJ uses algorithms and computer models to simulate how a building’s occupants will actually use a space. Their software links geometry with data to address specific problems such as the kinds of views available from different offices, and the best way to foster collaboration through visible sight lines.
According to Paul Audsley, who is NBBJ’s Director of Design Technology, design computation allows the development of “intelligent, flexible building models that provide instant visual feedback along with key supporting data to help us ‘prove’ our design concepts at the earliest possible stages.”
AI won’t replace human designers completely, but it will demand new skills and a mindset
capable of mastering algorithmic creativity. Whether it is a car, a chair or a HVAC system — future designers will leverage software to run through thousands of simulations before selecting a final solution.
Software makers are already experimenting with data-driven tools. Autodesk for example, have set up a research division called ‘Project Dreamcatcher’ to create a goal-directed design system. The system will allow designers to set certain parameters — such as material type and performance criteria — before evolutionary algorithms on a Cloud platform are given the opportunity to create thousands of valid design options and recommend the best-performing versions for further consideration.
Computational approaches to design have the potential to introduce truly disruptive ideas into our daily lives. If eliminating an iPhone headphone jack seemed crazy to you until this week, then you might also consider what other assumptions about our products, unimaginable to us now, are just an algorithmic permutation away from becoming a reality?