In response, Moncis, product manager of Google's search team, confirmed in an official blog recently that part of the audio collected was used to develop speech recognition systems to better adapt to various languages, accents and dialects.
"As part of our efforts to develop voice technology for more languages, we work with language experts around the world because they understand the nuances and accents of a particular language," said Monsis. These language experts review and transcribe a small amount of audio content to help us better understand these languages. This is a key part of the process of building voice technology, and it's also necessary to create products like Google Assistant.
Mencius also stressed that there were many safeguards throughout the process. In Google's comment system, these recordings will not be tagged into the user's account.
But because some audio files contain sensitive and recognizable information, VRT NWS can match some audio clips with users. Google said that users can access their privacy settings page to prevent Google from storing any audio.
"Making products for everyone is a core part of Google's DNA," he concluded. We adhere to high standards of privacy and security in product development and require our partners to adhere to the same standards.