AI design

All Hail the Extinction of the CAD Monkey

AI’s ability to take on repetitive tasks will give designers an opportunity to stay ahead of technology.

Ross Brady Ross Brady

Common wisdom holds that design work is impervious to automation because software can’t learn creativity, but that’s only partly true. Practical, repetitive tasks inherent in a design project are not only ripe for automation, but in many cases have already been ceded to artificial intelligence (AI). For young architects, this means a majority of the “CAD Monkey” work that has traditionally shaped early careers is about to become unnecessary.

If this makes you anxious, breathe easy — there’s some good news.

While removing repetitive, junior-level tasks from an architect’s workflow seems like it also removes young architects from the process, consider how it frees up time to investigate design options much deeper than before. Without grunt work, an architect’s role will shift from aesthetic technician to thoughtful curator. For young professionals, adapting to this change means learning to collaborate with artificial intelligence, communicate why design decisions are made, and facilitate multidisciplinary group work.

The prospect of artificial intelligence designing entire buildings is still a ways off, but the sophistication of current generative design tools may be shocking to someone who isn’t aware of them. A number of companies profiled recently in Architect already offer commercial applications of AI that generate every possible massing or layout option for a site or interior based on restrictions like local codes and physical conditions, as well as preferences like natural light and views.

In a test for a developer client that compared human designers to digital ones, humans “took an approach that was known to work based on past experience, but the algorithmic solutions moved beyond what you would typically think of,” resulting in “less obvious solutions that better balanced the developer’s need for profit against the community’s need for light, thermal comfort, energy conservation, and amenities.” The bottom line: designing in response to constraints can be taught to a machine, and the machine will inevitably produce better solutions than a person.

The rise of AI in architecture means the days of occupying interns with basic tasks like calculating stair treads are coming to an end — and that’s not a bad thing. Far from eliminating the need for junior designers, the use of AI fundamentally changes the approach an architect takes to the task of designing a building. The key to this is AI’s ability to generate designs a person would never have considered in the first place. Faced with solutions they couldn’t possibly conceive themselves, designers young an old may soon find themselves on equal footing.

On one hand, this shift favors seasoned architects who deeply understand the intricacies of the work being handed off to generative design tools. With newfound time to focus on feature items like elaborate geometry and centerpiece spaces, their years spent sweating the details can’t simply be replaced by cold, rational technology.

On the other hand, those with a fresh perspective and time to learn new skills can eventually outpace their predecessors by mastering the aspects of AI-assisted design that beg further exploration. As Sasha Crotty, Senior Product Manager for Revit recently stated to Geospatial World, “Automation is going to give architects more opportunity to investigate more design options and more time to consider the human factors that make up ‘good’ design.”

Collaborating with AI

If knowledge is power today, reasons will be power tomorrow. Understanding why certain AI-assisted design decisions are made and communicating these reasons to clients will soon be a top priority for architects. To reach that understanding, today’s architects will have to start treating generative design programs as legitimate collaborators instead of tools. This means digging deeply into why these new programs make the design decisions they make, becoming experts in their language and the patterns they follow.

An obvious way for architects to prepare for this shift is to become proficient in the basics of programming and code. Opportunities abound in this regard, from elective classes for current students to evening courses for working professionals. Beyond introductory coursework, however, an architect’s training provides invaluable insight into the mechanics of generative design software that even an expert programmer doesn’t have the context to discern.

Spacemaker, a company whose AI program optimizes building massing, sums up this value when describing one of its competencies as “quantifying and defining living space qualities,” a task that machine learning can easily best a person at, but only after a person defines what those qualities are. The architect of the near future will need to draw on their experience in order to determine those qualities for any particular project that only a trained designer would be able to conceive, though likely in a much more numerically analytical way than they do today.

Communicating Design with AI

via Pxhere

Communicating Design

If designing with AI requires understanding why it makes certain decisions, then communicating the reasons for those decisions to a client will become infinitely more important than if an architect were making them exclusively. While architects today may be able to bank on professional trust or evocative images to convince someone of a design’s value, speaking on behalf of a machine demands a very specific skill set.

Formal training in communications is already extremely beneficial to architects but rarely part of the profession’s education. Similar to learning code, a young architect seeking to correct this deficiency can easily find classes for working adults focused on public speaking or business communications, but for this skill, proper training must also be paired with consistent practice. Seeking out opportunities such as a local Toastmasters group, or even something as challenging as improv or stand up, can be good ways to develop this skill.

Facilitating Teamwork

With AI assuming the bulk of repetitive tasks in design work, project coordination will become even more crucial than it already is. Architects who leverage AI as part of a project’s workflow, successfully integrating it with the efforts of the numerous people required to design a building, will find themselves well-positioned against automation. Thus, the architect of tomorrow must devise the best processes for professionals with different backgrounds and training to work together.

Organizational dynamics is a well-established field, and a wealth of literature already exists for those willing to immerse themselves in it. Formal recognition, such as the PMP credential, can also raise a designer’s profile in this regard. Focused on facilitating collaboration and sorting out complex working methods, professional certification in this area can go a long way in helping architects adapt to technological change. Credentials aside, however, it’s important to remember that managing project workflows is a design problem – approaching this task with an architect’s sensibilities is an inherent advantage.

Guarding one’s self against the future isn’t easy, but with a little bit of personal effort and a flexible attitude, any of the strategies described here could become the secondary focus of an architect’s career. Faced with a shifting technological landscape, augmenting traditional design training and experience with complementary new skills is the best way forward.

Top image via Pxhere user Mohamed Hassan

+