Description

Microsoft has developed a new model of artificial intelligence for speech generation, which can come with many risks in the future, such as abuse, impersonation, and spoofing people's voices.

A new version of the text-to-speech AI model called VALL-E2 has been developed by Microsoft, which delivers fantastic performance in terms of the naturalness of the voice. Microsoft has introduced this AI model as a research project and has announced that it does not intend to make it publicly available to users due to the possibility of misuse of this AI model.

The VALL-E2 AI model is a new text-to-speech tool that can generate a sound quite similar to human speech in just a few seconds. According to Microsoft engineers, the model can produce accurate, natural, and human-comparable speech that can be mistaken for a real person's voice.

In introducing this model in their paper, Microsoft researchers said that the VALL-E2 model is the most advanced model among the neural child language models and is considered the latest development in this field. This model can also be considered a milestone in the transformation of text into speech, which has a completely equal function with humans. According to the researchers, the output quality of this model depends on the length and quality of the text command, and environmental factors such as background noise can also affect this quality.

Of course, Microsoft has stated that VALL-E2 AI is only a research project and that the technology is not going to be included in products or available to users. The company cited the potential risks of misuse of the technology as the reason for its non-release, saying that it is possible to impersonate a specific person with a fake voice with the help of this AI model.