The rise of artificial intelligence-generated audio statements poses major ethical and moral problems. To address this concern, a group of Arizona State University researchers have developed a prototype of a microphone that would authenticate voice recordings as human speech.
The microphone functions by identifying aspects of human speech in a recording. Then a watermark is added to flag the recording as authentic.
Visar Berisha is a professor with ASU’s engineering school and College of Health Solutions.
“Every time we speak, in order to produce this clear intelligible speech signal that we use to communicate with each other, we also produce a series of biosignals that the body emits," he said.
Visar says there are cases where authentication of human speech is important, such as protecting artist performances and likenesses from being duplicated without compensation. He also thinks the technology can enhance journalism, especially during election season when misinformation can spread.
As for other applications of the device, he says his team wants to expand their collaborative network with producers of mobile phone technology.