Amazon Alexa may one day be able to speak to you in the voice of deceased friends and family members.
According to Rohit Prasad, an Amazon senior vice president, the online retailer is working on a system that will allow Alexa to mimic any voice after hearing less than a minute of audio. Prasad stated that the goal is to "make the memories last" after "so many of us have lost people we love" during the pandemic.
Sky News says a video of the feature depicted a child who requested that their grandmother read them a story, and Alexa confirmed before changing her voice.
It's unclear how far the feature has progressed or when it might be available to Alexa voice assistants. The re: MARS (for machine learning, automation, robots, and space) event focuses on what Amazon is doing in ambient computing, including Alexa advancements, so we might not see this feature anytime soon.
There are also potential security concerns with Alexa's ability to precisely recreate a voice pattern, though we'll reserve judgment until we know how well Alexa can mimic a voice after only hearing it for a short period.
We'll also see how the feature is received; while it appears to require users to opt-in, there's an ethical concern about the rights of the deceased's voice and how long it'll be available.
The work delves into a field of technology that has been scrutinized for possible benefits and abuses. Microsoft Corp, for example, lately restricted which businesses could use its software to parrot voices. The goal is to assist people with communication disorders or other issues, but some are concerned that it could also be used to spread political deepfakes.
Amazon hopes that the project will help Alexa become more prevalent in the lives of shoppers. However, the public's attention has already transitioned elsewhere. An engineer at Alphabet Inc's Google made the contentious claim that a company chatbot had advanced to sentience.
Another Amazon executive stated on Tuesday that Alexa has 100 million customers worldwide, which corresponds to figures provided by the company for device sales since January 2019.
An Amazon spokesperson said that the voice-mimicking feature is not intended for deceased family members. It is based on recent advances in text-to-speech technology, as explained in a recent Amazon white paper, in which the team produced high-quality voice with far fewer data by using a voice filter rather than spending hours recording voice in a professional studio.
