Startup company Lyrebird is developing a technology which shows a scary future.
Given only one minute of source audio from a specific person Lyrebird’s technology can convert any text to the voice of that person.
Demonstration audio on their website here show what they are currently capable of, currently it does sound like processed audio, but it is only a matter of time before it will be hard to tell the artificial from the real. Combined with real time face mapping and we literally won’t be able to tell who actually said what.
The startup presents a range of uses for the technology, from speech synthesis for those who have lost their voice, one would assume permanently, to using celebrity voices to act as a personal assistant, or for use in video games. The lawyers are already queuing up I’m sure.
Lyrebird highlights some of the ethical issues arising from its technology.
“Voice recordings are currently considered as strong pieces of evidence in our societies and in particular in jurisdictions of many countries,” the company says. “Our technology questions the validity of such evidence, as it allows [someone] to easily manipulate audio recordings.”
Using voice as a form of bio-metric identification has just become obsolete.