The language experts review and transcribe a small set of queries to help Google better understand those languages. "This made it easy for us to find the people involved and confront them with the audio recordings". "And in order to understand the subtle differences and characteristics of the Dutch language, it still needs to learn a lot." notes VRT NWS adding that Google uses its online tool Crowdsource, in case the search engine has difficulty in analyzing a certain speech command.
Google responded to VRT NWS story in a blog post today by displaying zero accountability. When companies mess up like this, it becomes harder for them to defend themselves by saying "trust us" in future.
The subcommittee plans to "examine the impact of market power of online platforms on innovation and entrepreneurship", according to a release announcing the hearing.
However, VRT NWS shortlisted a few conversations and said that out of "more than a thousand excerpts", 153 conversations should never have been recorded because the user didn't say the "Ok Google" command.
After obtaining copies of some recordings, VRT NWS reached out to the users and had them verify their voice, or those of their children, talking to the digital assistant. Similar to Alexa, a recent report sheds light on how Google trains its virtual assistant by paying contractors to transcribe audio clips, including those where the personal assistant wasn't even invoked.
It seems Amazon isn't the only company who hires people to eavesdrop on conversations you have with your voice assistant - Google has now been caught out doing the same thing, too. Some of these recordings contained highly sensitive and private information.
The Register has asked the Belgian privacy watchdogs for comment, and we will update when we hear more.
"Okay Google. May I hit my wife?" and "Pornhub", are just a two of the thousands of phrases and conversation snippets recorded by Google and heard by the broadcaster, following a leak of over 1,000 recordings. "It doesn't take a rocket scientist to recover someone's identity; you simply have to listen carefully to what is being said".
In February, Google detailed that its Nest Guard, the centrepiece of the Nest Secure home alarm system, would soon receive Google Assistant functionality - meaning the device needed to have both a speaker and microphone.
In response Google admits to working with "language experts worldwide" to help improve their speech recognition.
"We just learned that one of [our] language reviewers has violated our data-security policies by leaking confidential Dutch audio data", Google said.
"We hold ourselves to high standards of privacy and security in product development, and hold our partners to these same standards", Google said.