search

WICG / speech-api

Web Speech API
https://wicg.github.io/speech-api/
145 stars 30 forks source link

Fingerprinting #46

Open foolip opened 5 years ago

foolip commented 5 years ago

See WebSpeech API mustn't allow fingerprinting. Issue moved here in order to discontinue use of Bugzilla for Speech API.

foolip commented 5 years ago

@marcoscaceres turns out there was an issue filed, please see the original bug thread too.

marcoscaceres commented 5 years ago

Thanks for the pointer.

bicknellr commented 2 years ago

This seems like it needs to be in the "Security and privacy considerations" section of the proposed spec text, but that section doesn't even mention it.

I don't think it makes sense that arbitrary sites be allowed to learn the set of installed voices without any user action. I've never seen a site that uses this API, let alone one that has convinced me that I should want it to both be able to access the complete list of voices and to override my chosen system default voice. What's the use case for this?

edit: To clarify, I'm questioning the use case for enumerating / selecting voices. I understand why you might want to ask the user's machine to read text to them as a general capability.