For the past two years, the world of design has been overwhelmed with the possibilities that AI suddenly brought into our lives. And while the rapid growth of AI capabilities is remarkable and sometimes scary, we should lean back a bit and ask ourselves a question: did it REALLY change the world of interior design as much as it feels like?
And the answer is "not really." And I'm not saying that just because I am an interior designer who wants to protect my profession for at least the next five years. Here are my arguments for it.
In the fundamental book "Space Planning Basics," Karlen and Fleming outlined the concept of the synthesis gap. Fundamentally, at all times—long before AI and even CAD systems—the gap between what a client wants and what they need was huge. Nowadays, with our relatively abundant lives and possibilities for building spaces, the gap has shrunk, but it still remains an everlasting problem for a professional designer.
While AI can already map out logical commonalities, it can't read the hidden message behind a client's request. When my client tells me, "I need to increase the number of workspaces in the existing office," they don't just want to cram more desks into the space, even in the most practical way. They need to consider changes to the whole workflow, people's roles, and new requirements that simply do not exist in AI data yet. Synthesis is a complex interaction of requirements and conditions that are always unique to each individual case, and it's just not possible to train AI on all of them because a new one emerges every day, making it impossible to predict its configuration.
AI can generate a hundred plans, but it can't help you actually choose the one to send to the contractor, bringing you right back to square one. It won't be able to convince you because it lacks the most important thing, which is human connection. We trust other humans' experiences more than anything else. What can AI know about someone's intolerance to AC or someone's need to sit near a window? What about privacy? Should the executives’ office be in isolation or be approachable by employees? Also, spatial programme is also a statement: should the office’s entrance be warm-toned and welcoming or vast and impressive? Do clients interact with employees in an open space, or do they meet a dedicated employee in an enclosed meeting area?
So far, AI often creates something impressive, but not what is needed.
So what can AI actually be used for right now? It does a decent job at eliminating routine tasks involving technical data—such as calculating material quantities, generating initial technical drafts, creating presentation graphics, automating relationship diagrams, and verifying technical details to eliminate human error.
These are deeply human nuances that remain unfathomable to artificial intelligence. While AI can pull up general spatial privacy requirements, it cannot understand the lived human experience of feeling uncomfortable during a phone call or having someone look over your shoulder. It is still a long way from that.
So far, AI often creates something impressive, but not what is needed.
A machine learning system's inability to comprehend the full complexity of human requirements stems from the nature of human development itself, which is shaped by countless unpredictable experiences that influence a person’s decisions.
It is safe to say that AI is here to stay, but with its own dedicated scope: replacing routine tasks and enhancing the overall design process within the boundaries of machine understanding. So far the gap remains open.
A machine learning system's inability to comprehend the full complexity of human requirements stems from the nature of human development itself, which is shaped by countless unpredictable experiences that influence a person’s decisions.
We will probably see a breakthrough someday that will shift AI to another level of understanding what is behind a human requirement, but until today, it remains a very prominently “computer-like” aid.