In addition to a new Echo speaker and a number of other announcements that Amazon has recently made, it has announced new capabilities for Alexa. He wants to make the digital assistant look more human.
Soon, you’ll be able to correct Alexa when she misunderstood you, saying “Alexa, that’s wrong!” or “Alexa, stop!”. When she realizes she made a mistake, the “nurse” will correct herself or ask a question to “fill in the gaps in her understanding.”
Amazon calls this feature “Teachable AI” and could help Alexa become smarter and stronger over time.
Learning capacity will first be on smart home devices in the coming months and will expand “into other areas in the future”. conform Amazon. It will be useful for situations where, for example, you have a preferred brightness setting for the bedroom lamp that Alexa doesn’t know yet. When you ask your assistant to turn on your light at your favorite level, Alexa can ask you what it is and remember for the future.
Alexa will sound more like a man, following the latest updates
Along with the learning function, Alexa will also sound more like a real person with pauses of breath, for example, when she talks to you. This is based on the “Neural Text-to-Speech” technology launched last year for more natural voices.
Alexa will adapt her answers based on the context of the conversation and adjust her tone by emphasizing certain words. Amazon calls this an “adaptation of the style of speech” and believes that these changes will “provide new and exciting experiences” when it begins launching later this year.
Amazon has also introduced a new feature called “Natural Turn Taking”, which is similar to Google Assistant’s ongoing conversations, but with more uses.
This will allow you to start a session where you can talk to Alexa without an activation word and talk to the assistant as if she were a person in your home. For example, after you say “Alexa, join my conversation,” your device becomes another member of the group discussion as you decide what to order for dinner.
Amazon says the wizard uses a combination of linguistic, visual, and acoustic cues to determine when someone is talking to Alexa, and then uses the context of what is being said to determine their response. So, if you watch the Echo Show and ask a question, Alexa will know that she should give an answer. Amazon said that this “transformation capacity” will reach the public next year.
Alexa, more friendly with children
Parents can also appreciate the new feature of children’s voice profiles, which will allow Alexa to move on to the “Kids Alexa” experience when she recognizes the little ones. This experience includes answers suitable for children, games, skills, books and music.
Those who already use Amazon Kids + will also see their children’s favorite premium skills and Audible books continuing the experience. A preview of the voice profiles will begin in the coming months, Amazon said, and will be available on all devices, not just the Echo Dot Kids Edition.
You can also ask Alexa to help you with the story. Amazon has introduced a feature called Reading Sidekick, which will allow the Assistant to read the accepted books in turn with a child, and listen to how well the little ones are doing. It will give encouragement and praise if your child is doing great and will help you if he is not doing well.
Alexa also comes to the aid of the elderly
The company also launches tools to help the elderly in your life. A new help center will allow you to set up a connection between two Alexa accounts, then view a high-level activity stream and set up alerts if your device doesn’t detect any activity at a certain time of day. Your older relatives may also designate you as an emergency contact who will be called immediately when they ask “Alexa, call for help!”.
Amazon takes your privacy a little more seriously and will give you more control over the voice data it keeps. Just like you can with Google Assistant, you can now ask Alexa about your privacy settings or tell her “Delete everything I said” at the end of this year. You can also choose the setting that does not save your voice recordings, which means that they will be deleted automatically after each request is processed.