Over the past few months, we have witnessed an incredible proliferation of AI-generated images across social media. As the novelty factor is slowly wearing off, it is time to talk about the impact on creative industries.
Recently, an artwork generated by Artificial Intelligence (AI) has won an art contest at the Colorado State Fair in Las Vegas, US. Entitled Théâtre D’opéra Spatial, Jason Allen’s work was produced using Midjourney, an AI image generator, and it was awarded the blue ribbon in a category open to emerging artists.
It promptly set the art world ablaze, splitting the crowd and dragging into the conversation a broad range of voices—from art pundits, designers to AI ethicists. While the aesthetic merit of the outcome can hardly be contested, the tug of war over intellectual authorship vs. technical ownership rages on.
This conundrum is not new. In the past, the invention of photography arguably contributed to the evolution of painting from a pursuit of faithful reproduction to one of creative interpretation. Rather than displacing artistic merit, a disruptive new medium could boost it into new directions.
Don’t miss: Property Trends: How AI, NFTs and Blockchain are Shaping Our Search for a Dream Home
As an architect and spatial designer, I can’t help but wonder how AI will impact my profession’s future. Before I elaborate further, I will make a voluntary admission: I am all for using AI. I started exploring Midjourney, an AI art generator, as soon as it became broadly available through Discord, and I found myself down the AI rabbit hole within a matter of minutes.
Visual communication is essential for designers, and here I was, able to generate with a few lines of text the most unpredictable—if not utterly unlikely—spatial environments. What could have taken hours or days to draw, model, and render mere months ago, now was being generated in front of my eyes in a matter of minutes. It all seemed suspiciously easy.
In case you missed it: What Beauty Means in the Metaverse Era
What could have taken hours or days to draw, model and render mere months ago, now was being generated in front of my eyes in a matter of minutes.
This instinctive hesitation to fully embrace AI as the future of design, was a healthy moment of pause as it made way for a more serious examination of my view on the creative process, the role of the designer, and ultimately the legitimacy and value of the creative outcome.
In a world where anyone can generate semi-plausible images, how can we practice discernment? For creatives, this can be a quantum leap and a litmus test.
Here’s what I have learned since, after a few more weeks of testing:
1. Great images and great design are two different things
Let’s face it: we all spend way more time scrolling our Instagram feed than we would like to admit. With great imagery on tap, we are probably living the most content-rich period in human history.
Beauty fatigue can be real, to the point where we conflate and confuse the visual quality of a picture with the actual merit of the design—whether spatial, functional, environmental, or experiential. Being superficially seduced by one’s own visual follies [beautiful but possibly impractical design] is a slippery slope that doesn’t invite learning from the outcome with some level of intent and objectivity.
2. Good design is all about the process
Designers are constantly tempted to chase the providential outcome. Great outcomes, however, are almost always the result of a thorough process.
Developing a critical ability to understand the rational iterative steps involved in formulating, testing and realising a vision is a strength that allows one to challenge and improve the final result. When working with AI, the purposeful testing of prompts is a great way to limit chaotic randomness and stay in control of the creative output.
In my Post-Nature series [created using the Midjourney AI-art generator], I have set myself the difficult challenge of describing and re-creating an atmosphere I know and live every day: the hot, wet, thick equatorial weather, and unique qualities of our tropical architecture, interiors and landscape. By setting clear evaluation criteria, I was able to objectively assess my images and work towards improving their ability to tell authentic biophilic stories.
3. Beyond the Pinterest effect
Over the past decade, Pinterest has become a creative crutch for most designers. “Relevance” algorithms have put the global design scene on a homogenous diet of reference images. Many design trends we have seen over the past few years—like arches, or soft pastels—are not accidental.
The AI’s computing power should be able to help us generate original “mood” imagery better and faster than compiling a Pinterest board. Occasionally, a surprise may come along.
In my case, the Construction Couture [also generated with Midjourney] collection started after a long site visit where I couldn’t help but wonder how construction site clothing—especially in the age of AR [augmented reality] and AI—could evolve to embrace new technologies, lighter materials, and sustainable fabrics.
Would a similar exploratory process allow designers to test, cross-pollinate and learn from other fields? I certainly hope so.
4. What's next
We are without a doubt living in the very early days of exploring AI as a design tool. As AI will get faster and more attuned to input, the human user element will also have to evolve and hone essential skills. If traditionally most designers spend a lot of time operating software and performing repetitive production tasks, once AI looks after that, what will be our next litmus test?
My belief is that conceptual clarity, ability to communicate our ideas and creative agility will once again shine—levelling the design playing field for the up-and-coming, the brave, and the articulate voices that may emerge from anywhere at this point. And this is something truly worth getting excited about.
Architect Razvan Ghilic-Micu is an extroverted natural communicator, and an architecture lead at Hassell Singapore. A global design voice with professional experience in Singapore, Shanghai, New York City and Toronto, Ghilic-Micu currently leads innovative mixed-use and commercial projects in Singapore and the region.
Outside of practice, he is active in the Singapore Institute of Architects as the chief editor of The Singapore Architect (TSA) Magazine, recently served as the festival director of the 2021 edition of Archifest, an architecture festival held annually in Singapore. Ghilic-Micu is frequently contributing to design publications and dialogues as a curator, panellist, and moderator.
