AI-Native Design
Your design system can express "primary button" but can't express "I'm 70% sure about this." It handles hover states but not confidence gradients. It ships a spinner but has no vocabulary for what "thinking" looks like when it carries meaning.
That's the gap. AI-native design is the practice of closing it — building design systems that treat AI behaviors as first-class primitives, not afterthoughts bolted onto component libraries built for a deterministic world.
Where to start
Start with the Framework — the eight shifts, the scoring methodology, the summary table. From there, the dimensions branch off. Trust Expression is the one most worth your time if you're building enterprise AI; Guardrails is the most forward-looking. The engine pages (Fluent UI, Material Design 3) score real systems against the framework. Rive is the argument for why design systems need a runtime layer.
This is a working section — writing continues as each dimension earns its full treatment. The contents nav is the current state of it.