Broadway

Complete News World

Microsoft can imitate any voice in the world

“Second Artificial Intelligence” is a sensation. With examples just three seconds long, an AI can imitate anyone’s voice — with significant stakes?

ad

The basics in brief

  • Microsoft and OpenAI offer the next AI software after GPT3 and DALL-E
  • VALL-E can create an audio imitation based on samples as short as three seconds long
  • Currently used for research purposes only, misuse would be destructive

After DALL-E comes VALL-E, and with it Microsoft’s next explosion in terms of AI. “AI seconds” are able to mimic sounds, tone of voice, and emotions. And to do so, it only needs to be trained in an original three-second sequence.

Microsoft and OpenAI are leaders in the development of artificial intelligence

VALL-E is similar to others from Microsoft Developed artificial intelligence tools such as chat and Dall-E. Are able to create images and texts based on keywords. Microsoft is one of the backers of OpenAI, the organization behind these tools.

VALL-E has been trained on 60,000 hours of English speech data, which enables it to create engaging and natural looking recordings. This is a major advancement in the field of AI-generated sound. Until now, it was impossible to create an accurate imitation of sounds with a very small sample size.

“AI seconds” is still not always accurate

However, it should be noted that the quality of the recordings produced by the Second AI may vary. Some are indistinguishable from the real sound and others sound tinny and are instantly recognizable. The depth of detail in a 3-second recording is very low. the most data are fed into artificial intelligence, the more accurate and realistic the results will be.

It is currently still in the non-public research stage

At this time, VALL-E is not yet available to the public, and it is not clear when it will be available. However, the potential uses for this tool are enormous. It can be used for voiceovers, virtual assistants, or even entertainment.

The way is paved for attempts to deceive and fraud

However, there are also concerns about Possibility of abuse, such as creating fake audio content. Disinformation can spread or fraud attempts can begin. Therefore, ethical and legal issues must be considered when developing and issuing such a tool.

More on this topic:

Microsoft data