Traditional UI/UX patterns (buttons, forms, ...) are no longer enough. AI systems perceive, reason, and act, we need a new design language that reflects uncertainty, autonomy, and context. This talk introduces a pattern language for AI-driven interfaces, exploring how we’ll design when users don’t just click-but converse, correct, branch, and explore over time and space. We’ll look at how AI transforms the fundamentals of interaction - how intent is captured, how systems explain themselves, and how users build trust in dynamic, adaptive environments. It’s an invitation to rethink the interface as a living dialogue between human and machine - one that learns, adapts, and reasons alongside us.