Tag Archives: TOOLS

#171: AI: BEWARE AND BEWILDERED

June 16, 2023

AI design by Manas Bhatia.

Do we need architects to create architecture? With artificial intelligence (“AI”) the answer is yes and no.

AI design by Tim Fu.

Years ago when writing my first book, I stumbled upon an AI app that could, with a click of the mouse, re-write my chapters in the “style of Hemmingway” or “style of Faulkner.” The results were not just entertaining, but convincing. But AI could only re-write my book, not write it from scratch. At least not back then.

Fast forward to today. AI can author screenplays, poetry, and entire novels. The question of whether the work is worthwhile remains the question, as AI delivers increasingly better results every day.

AI design by Anna Fixsen.

AI has been used in car design, healthcare, and manufacturing. Why not architecture? A decade ago, when AI invaded our creative turf, we responded defensively, “The AI results aren’t good at all” or “AI can’t replace the human hand and the personal touch I have with my clients.” But such reactions are shifting as the more open-minded see AI has yet another powerful tool to augment the work we do—tools like a T-square, AutoCAD, BIM, and 3D printing. So don’t worry: AI will replace you only if you let it.

AI image from illustrarch.com.

As glamorous and exciting as the design process is, much of architecture is analytical problem solving. With architecture as part science, AI can be ideal for analysis. Let the software perform code research, square footage analysis, cost estimating, quantity tracking, energy modeling, and parking counts.

AI design by Stephen Coorlas.

Regarding the creative process, AI-software requires a facilitator, someone to prompt the program. If you ask AI to design an office space, the result may be a boring office with low ceilings and generic furniture. But if you ask AI to design “a creative office space with cathedral ceilings and Italian work stations,” the resulting design will be more inspired.

AI design by Stephen Coorlas.

But are any of these results actually good? Sure, AI has speed and the capacity to generate options, but its explorations into new shapes and cinematic atmosphere, seems more like stage sets for a science fantasy flick than a livable engaging work of architecture—contrived and extreme vs. authentic and grounded.

Théâtre D’opéra Spatial by AI, First Place Winner, Colorado State Fair Fine Art Show.

At last year’s Colorado State Fair Fine Art Show, a controversy of AI made national headlines. For the digital art category, digital photographers/artists proudly submitted their pieces—painstakingly curated and fetishized. Yet an AI-generated work entitled, Théâtre D’opéra Spatial, took first place. But if there is no artist, shouldn’t this work be disqualified? No human hand was responsible for this striking work of art. Should a computer and its software be eligible to compete? Jason Allen, the “winning” programmer, came forward with no pretenses of having been the artist or author. Despite many questions and debates, the first prize ribbon stood, and a new controversial world of authorship has begun.

AI design by Andrew Kudless, Matsys Design.

What are the ethics surrounding AI? Is there a morality to how and when AI should be used? As architects, we have a responsibility to “protect the health, safety, and welfare of the occupants.” In fact, we are licensed by the state to uphold our responsibilities, and held liable if we fail. Imagine designing a movie theater without the proper exits—and hundreds die in a fire.

So who is responsible for an AI-generated building design? The AI process is not flawless nor neutral, as one would hope science and technology to be. Also, AI lacks transparency. The machinations of AI are not comprehensible to us humans, as there might be biases, prejudices, and stereotypes. Who is accountable? Only our murky future holds the answers.

Me recreated by Lensa AI.

#103: LIMITED BY THE TOOLS OF THE TRADE

July 12, 2019

Old days of architectural drafting (photo from Archinect)

For most architects, the design starts inside the brain. We are then challenged to extract that creative spark out of our head and on to paper, or these days, on to a computer screen. Urgently, we grasps at the tools of our trade to convert the abstract ideas into some visual form of communication, i.e., the sketch on the back of an envelope, the first computer drawing, or the crude paper model.

Often, our ideas are grander, more ambitious, than any tool can capture. Tools have limits, whereas our artistic spirits do not.

T-square and triangle (drawing from etc.usf.edu)

The old days of architecture embraced simple non-mechanical tools, such as the T-square and the triangle. This allowed us to merely draw parallel lines and only four angles—30, 45, 60 and 90 degrees. If our brain generated an architectural idea with a curved shape or at an 18.5 degree angle, our tools were challenged to capture the idea.

A new tool came along: the adjustable triangle. No longer a static piece of wood or plastic—this tool was mechanical, moving upon its little hinge. The adjustable triangle freed the architect to now make any angle of choice. During school, we used to joke by pointing out when students purchased his/her first adjustable triangle, because  their drawings all of a sudden had a new complexity of diagonal lines.

The adjustable triangle (photo by Anthony Poon)
Lost tools of the trade (photo by Anthony Poon)

Alongside other instruments such as the compass, French curves, elliptical templates, etc., new ideas could be expressed. Architecture started to have move diagonals, more curves, more complexity. Again, we poked fun, “With these new house designs and the angles, Frank Lloyd Wright must have purchased fancier drawings tools for his staff!”

Floor plan of Frank Lloyd Wright’s Taliesin West, Scottsdale, Arizona

A quantum leap in communicating design ideas arrived with digital technology. These days, almost anything architects can dream up can be captured using today’s devices. With algorithms, computers are not just communicating ideas that are in our heads, but are generating ideas without our heads.

Using a parametric algorithm with the software, Grasshopper, to design a trellis structure in South Pasadena, California, by Poon Design. The material is polyethylene panels, the same plastic as kitchen cutting boards—used to express the home owner’s passion for cooking (photo by Sharon Yang)

Here is the question: just because we can think it, just because we can draw it, just because we can build it—should we? Just because software can describe a heroic complex form (like CATIA for Frank Gehry), just because a computer can document a complex pattern for water jet cutting a sheet of steel, just because 3ds Max and Maxwell Render can produce a near photographically realistic image, should we have technology replace the use of our brains and our hands?

Taiyan Museum of Art, China (photo from imagenesmy.com)
Heydar Aliyev Center, Baku, Azerbaijan (photo from igsmag.com)

One example: If you tour an architecture school or many architecture design studios, you will see the excessive use of the 3d printer. With limited time on the computer and a few clicks of the mouse, dozens of physical models of a particular design theme are produced in plastic. I argue that most of these variations-on-a-theme are insignificant. Just because an architect can generate 20 similar ideas, doesn’t mean that all these ideas have merit. Wouldn’t it be better to develop one idea carefully, strategically and thoughtfully?

My personal preference is to design ideas that are more hand crafted, then machine produced—relying more on heart than tools.

Garden lights using handwoven baskets from Ten Thousand Villages, a nonprofit fair trade organization sourcing from Africa. The light source is in the ground shining up into the basket, providing a soft downward glow, by Poon Design (photo by Poon Design)
© Poon Design Inc.