"Humans will over time become even more vulnerable to such attacks," warns Dr. Nitesh Saxena of UAB's Department of Computer Science.

New Scientist reports that Baidu, a Chinese technology company focused on internet-related services and products, has announced that it can copy someone's voice using less than a minute of audio of the person talking. This technology could allow those who have lost the ability to speak to create a personalized digital voice or to personalize technology like games, toys, GPS, and other tools that incorporate voices.

It could also be used to break into voice authentication systems. Four years ago, Dr. Nitesh Saxena of the Department of Computer Science showed that existing technology could already fool these systems 80-90 percent of the time. And while authentication systems may be improved to guard against synthesized voices, humans may not be able to tell the difference between real and synthesized voices.

"Humans will over time become even more vulnerable to such attacks," Saxena told New Scientist, warning that voice synthesis could be used to commit fraud or create fake news by doctoring a politician's speech.

Read more about Saxena's research on voice imitation at UAB News.

Science in Motion

Back to Top