THE LIMITS OF ARTIFICIAL INTELLIGENCE IN DESIGN

“Discussions around artificial intelligence are often marked by an unusual degree of certainty. Predictions oscillate between utopia and collapse, efficiency and extinction. Yet beneath this confidence lies a contradiction. No one can accurately predict how artificial intelligence will reshape creative and professional practices, particularly in design. The speed of its development gives the impression of inevitability, but inevitability is not the same as clarity.

Part of this confusion stems from the language used to describe it. Intelligence suggests agency, intention and independent thought. Artificial intelligence possesses none of these. It does not think, reflect or understand in the human sense. It processes information, identifies patterns and recombines what it has already encountered. Its outputs are not ideas formed through intent, but results generated through probability.

That distinction, however, is less reassuring than it might seem. Human creativity operates in a remarkably similar way. People absorb information throughout their lives, consciously and unconsciously, and recombine it into new configurations. From ideas to products, from personal beliefs to political positions, from poems to songs, creation is almost always a process of recombination. It is shaped through accumulation, association and reinterpretation. In this sense, artificial intelligence mirrors a fundamental aspect of human behaviour rather than opposing it.

This is precisely why its impact feels so disruptive. Artificial intelligence accelerates processes that were previously slow and personal. It assists professionals by reducing friction, automating repetitive tasks and compressing time. At the same time, it displaces roles that relied on those same processes, including roles traditionally considered creative. What was once a skill becomes a pattern. What was once judgment becomes an algorithm.

The consequences are already visible in digital environments. Websites and applications increasingly resemble one another. Artificial intelligence is highly effective at identifying what performs well and reproducing it at scale. Layouts converge, interactions stabilise and visual language becomes predictable. Optimisation produces consistency, but consistency erodes distinction.

This is where design risks losing something essential. Imperfection is not a flaw to be eliminated, but a condition of variation. Irregularity produces character. Ambiguity invites interpretation. The world is not uniform because people are not uniform. When systems are optimised exclusively around recognition and replication, they flatten difference in favour of efficiency.

There is also a more fundamental limitation. Artificial intelligence does not encounter objects in the world. It has no physical presence, no sensory feedback, no embodied experience. When faced with a new design, it attempts to classify it using references it already knows. It approximates understanding through resemblance. What it lacks is refinement. It cannot fully grasp how an object behaves once it leaves the screen and enters daily life.

This distinction matters particularly in product design. A product does not exist solely as an image or a concept. Once conceived, it must perform in the physical world, among people, habits and contexts that are rarely predictable. Materials age, users adapt, environments interfere. These variables resist abstraction. They require observation, adjustment and judgment. They demand a sensitivity that extends beyond pattern recognition.

For this reason, artificial intelligence occupies an ambiguous position in design. It is an extremely powerful tool during conceptualisation, analysis and optimisation. Yet the moment a product is released into reality, its relevance diminishes. The world does not behave like data. It responds to nuance, misuse and contradiction.

Much has been said about artificial intelligence and creativity. Far less is said about its limits in relation to objects that must exist beyond systems. Design continues beyond the conceptual phase, shaped by factors that only become visible once an object enters the world.”

– Mike Simonelli

January 2025