Amazon is working on a new controversial and quite creepy AI feature for its Alexa devices and future voice assistants. The company demonstrated its new voice generation technology on its smart speaker during its MARS conference at which Alexa was able to mimic a deceased person through voice.
- Amazon announces an experimental AI voice feature.
- The voice mimicry enables Alexa to generate voices of any person.
- Amazon did not mention if this will really be a feature in its future smart devices.
While video and audio deepfakes are starting to become prominent, the technology behind it is still somewhat limited to high-end computers or expensive software solutions. That is likely to change in the coming years as more companies are introducing AIs capable of such functions to general consumers.
Similarly, Amazon has shown the potential of what its AI technology can do via its Alexa assistant. At the event, a video was shown of a boy behind the camera commanding Alexa to continue reading the Wizard of Oz book through his deceased grandma’s voice. This is somewhat concerning and impressive at the same time.
In addition, the company says that its technology also works with any voice using less than a minute of recording input. For example, anyone can generate a famous person’s voice as long as they have a recorded audio clip. There’s also a possibility that users can replace Alexa’s default vocal sound in the future.
Amazon did not confirm if this experimental mimicry feature will eventually come to its smart speakers. However, we will not be surprised if this will become a key function for Alexa users considering more companies are racing to develop intelligent machines and assistants.
How would you use voice mimicry or audio deepfakes in the future? Hit us up in the comment section.