Google is defending its practice of letting human workers located around the world to listen to audio recordings of conversations between Google Assistant software and users. Most of these employees happen to be contract workers from various parts of the world. The response comes after revelations from Belgian public broadcaster VRT NWS explained how contract workers sometimes listen to sensitive audio captured by Google Assistant accidentally.
Google in a blog post said, that it is going to take precautions in order to protect user identities and that it has “A number of protections in place to prevent” the so-called false accepts which is when activated by Google Assistant on a device such as Google home speaker without the wake word being intentionally used by the user.
The company also stated that its employees review these conversations in order to help Google’s software operate in multiple languages. David Monsees, a product manager on the Google Search team who authored the blog post writes:
“This is a critical part of the process of building speech technology and it necessary to creating products like the Google Assistant.”
He also added:
“We just learned that one of these language reviewers has violated our data security policies by leaking confidential Dutch audio data.”
Monsees referred to the snippets of audio the Belgian contract worker shared with VRT NWS.
“Our security and privacy response teams have been activated on this issue, are investigating, and we all take action. We are conducting a full review of our safeguards in this space to prevent misconduct like this from happening again.”
Additionally, Google also claimed that just 0.2% of all audio snippets are reviewed by language experts. Monsees says:
“Audio snippets are not associated with user accounts as part of the review process, and reviewers are directed not to transcribe background conversations or other noises, and only to transcribe snippets that are directed to Google.”
“We are always working to improve how we explain our settings and privacy practices to people and will be reviewing opportunities to further clarify how data is used to improve speech technology.”