Google Launches Database to Prevent AI Manipulation by Fake Audio

Google has created an enormous database that has thousands of hours of fake audio in a wide variety of accents. This is being done to help prevent its artificial intelligence (AI) from being manipulated or fooled by fake audio recordings. It is heartening to see Google is putting so much effort into maintaining a sense of reality in a world that is hyper real as the internet, where a lot of fake news ends up having an enormous impact on the world even if there are no sources to back it up.

This database is also important because of the fact that it marks Google fulfilling a promise that it initially made back when it launched the Google News Initiative. Congress has taken initiative against AI manipulation by false audio recordings, and this has probably contributed to Google taking this problem seriously over the past few months. However, one thing that should be kept in mind is that this database has not been tested yet. It could very well end up failing, but only time will tell whether or not Google has managed to do a good job at preserving the security of users on the internet.
"As we published in our AI Principles last year, we take seriously our responsibility both to engage with the external research community, and to apply strong safety practices to avoid unintended results that create risks of harm.", explained Daisy Stanton, Google AI Software Engineer, in a blog post. Adding further, "We're also firmly committed to Google News Initiative's charter to help journalism thrive in the digital age, and our support for the ASVspoof challenge is an important step along the way."
Google releases dataset to help AI systems spot fake audio recordings
Photo: Khari Johnson / VentureBeat

Read Next: These Random Pattern Graphics in Google Image Search Results Are One Of The Weirdest Things On The Internet
Previous Post Next Post