You can automate bad design—but it’s still bad design.

Lately, I’ve seen a growing assumption that integrating AI into a product will automatically fix its user experience. It won’t.

AI can personalize, predict, and accelerate workflows — but it can’t disguise poor design decisions, inconsistent navigation, or confusing interfaces. If the fundamentals of usability are broken, AI just adds complexity on top of chaos.

Dr. Jakob Nielsen’s 10 Usability Heuristics are as relevant now as ever. They’re timeless because they’re about human cognition, not technology. And if your product violates them, no amount of AI will save it.

Let’s recap a few that I see AI breaking already:

• Visibility of System Status: When AI makes decisions without clear feedback, users lose trust.
• User Control and Freedom: Autocorrects, auto-fills, and “smart” suggestions that override user intent cause frustration, not delight.
• Error Prevention: Machine learning models can introduce new types of errors — ones humans don’t understand how to fix.
• Consistency and Standards: AI-generated UIs that shift tone or behavior between screens create cognitive dissonance.

AI is powerful, but only when it’s built on a solid foundation of usability. The smartest systems in the world still need clarity, predictability, and respect for user intent.

As designers, our role is to ensure AI enhances the human experience, not obscures it. Before adding “AI-powered” anything, make sure your experience already honors Nielsen’s principles — because bad UX, automated, is still bad UX.

💬 What examples have you seen where AI made an experience worse instead of better?