As bizarre as it may sound, it is not science fiction. After community projects that mimic some iconic voices, such as Morgan Freeman’s, this time Amazon itself announced that it was working on a new artificial intelligence project to customize Alex, with which we can imitate the voice of any person.

The bizarre revelation was made at re: Mars’ annual conference, where Amazon’s senior vice president and chief scientist Alexa Rohit Prasad described the series in detail. new features will come to smart assistant in the future. Undoubtedly, the most noticeable announcement was this small demo, in which they showed Alex reading a story to her child at bedtime, using the voice of his deceased grandmother.

The miracle that was achieved with just one minute sound of a woman as well Prasad himself explained during the conference: «it required inventions in which we had to learn produce high quality voice with less than a minute of recording time compared to recording hours in the studio. The way we made it possible was to frame the problem as a speech conversion task and not a speech generation path.«.

At this point, however, it seems to be a demonstration of his abilities rather than a feature we will soon see available to the voice assistant. Not only has Amazon not given any details about its possible date, but it has also avoided talking about general availability. However, it is clear that this functionality is very attractive, so it remains to be seen how Amazon will proceed with its commercialization.

with the proviso that We will need to see the restrictions set by the company for this featurebecause it could lead to certain problems in the face of identity theft. Even so, it is certainly inspiring to see the great advances that artificial intelligence continues to offer and overcome the barriers that we considered virtually unattainable to design a future full of possibilities.