Fact · Dictation on Mac
Apple's 2017 Hey Siri article says the detector uses a Deep Neural Network to turn the acoustic pattern of your voice into probabilities over speech sounds.
Checked against its source on October 16, 2026
In the source's own words
“uses a Deep Neural Network (DNN) to convert the acoustic pattern of your voice at each instant into a probability distribution over speech sounds”
Hey Siri: An On-device DNN-powered Voice Trigger for Apple's Personal Assistant, Apple Machine Learning Research, 2017, Opening section on how the detector works.
Open the original- What it measures
- The article describes the wake-word detector for 'Hey Siri', which runs on the device. It shows how Apple's speech stack used neural networks in 2017. It does not describe the Dictation key.
- What it does not show
- A 2017 design for a wake-word detector, not a description of Dictation's current models, which Apple has not published in the same detail.