Siri Dictation Gets Complete Rebuild With Auto Punctuation

The real story in Apple’s Siri update is not the voice customization sliders. It’s that the dictation engine is finally being rebuilt from the ground up, with automatic punctuation, capitalization, and real-time formatting handled without user intervention.

Dictation on Apple devices has long been a quiet embarrassment: functional enough to avoid complaints, unreliable enough that most people abandoned it after one bad experience. The promise here is that users can speak naturally and get polished text back, which is the baseline expectation Android users have had from Google’s voice input for years.

The voice customization piece is more incremental. Two sliders (expressiveness and pace) let users tune how Siri sounds, and Apple says voices will be more expressive on devices running its most advanced on-device model. As of the first developer beta, only an American English voice is available, which suggests the rollout will be staged and other locales are pending.

Hardware Dependency

The “most advanced on-device model” qualifier is doing a lot of work in Apple’s announcement. It almost certainly means these features will be gated to recent iPhone and iPad hardware, continuing a pattern where Apple’s AI improvements effectively deprecate devices that are only a few years old.

The combination of better dictation and voice customization points at something Apple has not said directly: Siri is being repositioned as an interface layer that people might actually want to talk to, rather than tap around. Whether the dictation accuracy holds up outside of demo conditions is the only thing that matters now.