In an era when algorithms can compose symphonies, design furniture, and even mimic architects’ sketches, the question isn’t what AI can create—but what it can’t feel. And that’s where wit steps in. The moment when a chair makes you smile or a building makes you pause and think, “Ah, that’s brilliant”, it’s a micro-exchange between creator and observer, full of nuance, timing, and intent.
AI, however, doesn’t yet intend anything. It doesn’t laugh, cringe, or feel the delightful tension between the expected and the absurd. Its output can be humorous in appearance, but not in origin. Wit requires awareness: an understanding of context, irony, and the human condition. It emerges from our ability to perceive contradiction and find beauty in it.[1][4]
AI systems can mimic humor through patterns, but not experience it. They recognise correlations, not contradictions.[1][5]
They can detect what people have laughed at before, but they can’t sense why something is funny—or when not to be.[2][3]
When a designer adds wit, they reveal a pulse—a human context behind form and function. Wit signals that someone, somewhere, was thinking not just about utility, but about you. It’s the human fingerprint in design separating intelligence from understanding: emotion translated into form.
Image: created with AI
[1] AI Humor Generation: Cognitive, Social and Creative Skills for Effective Humor
[2] Can Robots Crack a Joke? The Limits of AI’s Humor Understanding
[3] Why AI Can’t Write Jokes: Google DeepMind’s Study Revealed
[4] Limitations of AI humor | Dotneteers.net
[5] How funny is ChatGPT? A comparison of human- and A.I.-produced jokes
THE ARCHITECTS OF WIT 