Allow Automatic Update of Speech Data
This device-wide policy setting determines whether Windows may receive updated speech recognition and speech synthesis models.
since 0.9.8.11
Purpose
Show registry keys (1)
HKLM\Software\Policies\Microsoft\Speech\AllowSpeechModelUpdate
This device-wide policy setting determines whether Windows may receive updated speech recognition and speech synthesis models. Speech models are non-executable data files used by the speech engine to convert audio to text or text to speech.
A value of 1 allows model updates. Windows then checks periodically for updated models and downloads available data from a Microsoft service by using the Background Intelligent Transfer Service (BITS). A value of 0 prevents these model updates. Microsoft documents 1 as the default value.
Notes
Additional effects
Updated models can improve speech-processing accuracy and performance. When updates are not allowed, the models already present on the device generally remain available, but they do not receive subsequent improvements through this update mechanism.
Security
Allowing updates permits periodic network connections to a Microsoft service and the download of new model data. Microsoft identifies the downloaded speech models as non-executable data files.
This setting controls only speech-model updates. It does not determine whether online speech recognition may be used or whether voice input is sent to Microsoft speech services.
Notes
AllowSpeechModelUpdate is a device policy backed by the Speech.admx administrative template.
Microsoft documentation
Microsoft Learn: Policy CSP – Speech