Fact · Voice privacy and security
Apple says that on-device machine learning benefits privacy because the data for inference workloads stays on the device rather than on a server.
Checked against its source on October 16, 2026
In the source's own words
“Increasing the adoption of on-device ML deployment will also benefit user privacy, since data for inference workloads remains on-device, not on the server.”
Deploying Transformers on the Apple Neural Engine, Apple Machine Learning Research, Apple Machine Learning Research, 2022, Introduction, Apple Neural Engine section.
Open the original- What it measures
- Apple's machine learning team explaining why it runs Transformer neural networks on the Apple Neural Engine in Apple silicon.
- What it does not show
- A general principle about Apple's on-device machine learning. It does not say which Dictation requests run on the device, which Keyboard settings and Apple's privacy page answer for each language.