Google confirms leak of audio data from Google Assistant

Google has confirmed that confidential Dutch audio data has been leaked by a language reviewer. In a blog post, David Monsees, product manager, search said as part of developing speech technology for more languages, Google partners with language experts around the world who review and transcribe queries to help the Google team better understand those languages.

He added that one of these language reviewers violated the data security policies by leaking confidential audio data from Google Assistant. Investigations are currently ongoing and the company is also conducting a full review of its safeguards to prevent misconduct such as this from happening again.

According to CNBC, over 1,000 audio recordings of customer conversations with the Google Assistant were released to a Belgian news outlet. The news site, VRT, also said it was able to identify several individuals in the conversations, as well as personal details such as addresses.

However, Monsees said in the blog post that language experts only review around 0.2% of all audio snippets, and that these audio snippets are not associated with user accounts. Moreover, Google claims that reviewers are directed not to transcribe background conversations or other noises, and only to transcribe snippets that are directed to the company itself.

The Google Assistant only sends audio to Google after the device detects interaction with the Assistant, such as by saying “Hey Google” or by physically triggering the Google Assistant. Monsees also explained that some noise or words in the background might potentially be picked up and interpreted by the software, such as a hotword "Ok Google". He added that there are protections in place to prevent these sort of "false accepts". According to Monsees, Google is willing to provide users with tools to manage and control the data stored in accounts.

"Building products for everyone is a core part of our DNA at Google. We hold ourselves to high standards of privacy and security in product development, and hold our partners to these same standards. We’re 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," he said.

(Photo courtesy: 123RF)