At its annual MARS conference, Amazon introduces a new feature for your devices Alexawhich can use the artificial intelligence of its software to imitate the voices of deceased people, an option that is too specific, a little strange, and which will undoubtedly generate mixed opinions as it becomes more popular.
voice from the grave
In his presentation, Amazon demonstrated this new option by showing a video in which a child asks a device Alexa read him a bedtime story in the voice of his deceased grandmother. As we mentioned, it’s a little scary, but Alexa says it can help you get through the death of a family member, especially in times of crisis like the one humanity just went through.
“As you have seen in this experience, instead of being the voice of Alexa reading the book, it’s his grandmother’s voice,” he said. Rohid Prasadchief scientist for artificial intelligence Alexawho also mentioned that adding “human attributes” to information systems AI it was important “in these times of a pandemic when many of us have lost a loved one. While the AI can’t take away that pain of loss, it can really make the memories last.”
According to the expert, the systems of the AI Alexa they can learn to imitate a person’s voice after analyzing a single minute of audio recording. So in theory, if you have a voice memo of your grandmother, or even a video of her talking, that might be enough for the Artificial intelligence can create a new voice that looks like it. However, many doubts remain about this technology and you would have to see it act in real time to know how faithful it can be. Furthermore, it raises questions as to whether the AI can imitate anyone and if it can be a risk for them. Likewise, it can be used for negative purposes if, for example, the voice is not reading a bedtime story but something else that could mislead loved ones.
It is possible that for this reason Amazon he did not say if he would ever make this technology public. In the presentation, only this feature was shown but it was not mentioned if it would come to a future device Alexa.
The truth is that the technology is a natural advance considering other similar ones that are already responsible for mimicking voices, which is known as counterfeit audio and they’ve already been used in movies, TV, video games, podcasts, among other media. In addition, there are already very simple tools with which some people can listen to any text with the voice of a fictional character. Many of them still need to be refined to be true to the original voices, so it might be a while before we see them become more mainstream.