Blog: Android

Speech recognition mobile assistants, the lowdown starts here

Antonio Cassidy 02 Mar 2015

voice-command

Given our previous post about the Samsung Voyeur TV it got us thinking about other voice based assistants. In days past all translation of voice to commands was performed locally on the device, however these days it looks like using a web based service is the preferred method.

 

We looked at the Terms of Service for the big 4 platforms to see what we were agreeing to, worryingly Blackberry admits to send some data in plaintext!

Over the coming weeks we will be tearing these assistants apart to view how the data is being transmitted and what’s being sent.

Blackberry Assistant

When you use BlackBerry Assistant, your voice instructions (“Audio Data”) will be recorded and used to process your requests and to provide and improve the products and services available to you. In order to accurately recognize and process your requests, Audio Data, information about your contacts (“Contacts”), the names of applications and songs in your music playlist(s), location information (where available), and device-related information (collectively “User Data”) may be sent to servers of BlackBerry Limited or its subsidiary companies and service providers (“BlackBerry”), and is subject to BlackBerry’s Privacy Policy available at http://www.blackberry.com/legal.
Your Contacts will be sent to BlackBerry in encrypted form, but for performance reasons other User Data and Responses may be transmitted in an unencrypted format.

Windows Mobile: Cortana

To help Cortana understand the way you speak and better understand your voice commands, Cortana sends speech data to Microsoft to build personalized speech models and improve speech recognition for Cortana and other Microsoft products and services. Speech data sent to Microsoft includes voice recordings, as well as associated performance data (such as changes you manually make to text and the final text result). Cortana also collects and uses other information, such as the names of your contacts, how often you call them, the titles of your calendar events, and words you’ve added to the dictionary.
Information that is collected by or sent to Microsoft by Windows Phone may be stored and processed in the United States or any other country in which Microsoft or its affiliates, subsidiaries, or service providers maintain facilities. Microsoft abides by the U.S.-EU Safe Harbor Framework and the U.S.-Swiss Safe Harbor Framework as set forth by the U.S. Department of Commerce regarding the collection, use, and retention of data from the European Economic Area, and Switzerland. To learn more about the Safe Harbor program, and to view our certification, please visit http://www.export.gov/safeharbor/.

iOS: Siri

When you use Siri or Dictation, the things you say will be recorded and sent to Apple in order to convert what you say into text and to process your requests. Your device will also send Apple other information, such as your name and nickname; the names, nicknames, and relationship with you (e.g., “my dad”) of your address book contacts; song names in your collection, and HomeKit-enabled devices in your home (e.g., “livingroom lights”) (collectively, your “User Data”). All of this data is used to help Siri and Dictation understand you better and recognize what you say. It is not linked to other data that Apple may have from your use of other Apple services. By using Siri or Dictation, you agree and consent to Apple’s and its subsidiaries’ and agents’ transmission, collection, maintenance, processing, and use of this information, including your voice input and User Data, to provide and improve Siri, Dictation, and dictation functionality in other Apple products and services.

Android: Google Now

Google’s Privacy Statement and Terms of Service do not specifically mention the google now service nor does it mention any dictation or voice transport. Like the other providers use of the services grants google permissions to share your data with third parties. Worryingly the volume and type of data which google aggregate to supply tailored notifications is somewhat stalkerish!