Alexa will have her own opinions on beer, TV shows and more

Alexa will have her own opinions on beer, TV shows and more
From TechCrunch - January 12, 2018

Alexa is starting to think for herself. The personal assistant thats built into Amazon devices like Echo smart speakers and Fire TV may better distinguish itself from competitors like Google Assistant in the future by offering up not just facts or recommendations, but by having her own opinions on things that werent explicitly added to her programming.

This will come into play, in particular, in the living room and in home entertainment scenarios, where Alexa will start to offer consumers more than just information on whats popular on TV tonight, for example, or what she thinks you might like to watch. She will also be able to offer what she personally recommends.

The idea is that speaking to an assistant shouldnt be about performing fact-based queries via voice, but having a conversation,explained Amazon Fire TV VP Marc Whitten, in a conversation at CES. That means Alexa will have to feel less like a voice search engine and more like someone you can really talk to.

Having an opinion makes you more interesting, even as an assistant, he said.

The Fire TV team has already seen the differences in how consumers act when talking into a voice remote by pressing a button, and talking to Alexa hands-free. In the former case, its like people are dictating queries into a search box.But when speaking hands-free, they start to ask more nuanced questions, like show me comedies between 88 and 92, Whitten noted.

This naturally leads people to ask a wider variety of questions, like whats interesting tonight? or what should I watch?, which Alexa could respond to using data on TV shows popularity, or things she already knows about your preferences and viewing behavior in order to make personal recommendations.

But Alexas opinions would be neither of those things. They would be her own.

This is the 2018 version of the video buff at the video rental store, said Whitten. Its like when you ask a friend whats good to watch, and they offer a suggestion thats not necessarily one of the top-rated showsjust something they are enjoying.

This is the power of machine learning. One of the most interesting things were going at is how do you design an assistant that feels like youre having a conversation with someone, Whitten said.

Machine learning and deep learning networks are a key part of how Alexa will eventually offer more than just the fact graphone of Googles key strengths today. A separate team at Amazon has been rapidly improving Alexas ability to answer questions by adding more fact-based information to Alexas knowledge base, while also observing where holes still exist by analyzing users queries that the assistant cant answer.

But while facts are important, what makes an assistant feel fun and friendly are its quirks, jokes, and other tricks that give it a personality.


Continue reading at TechCrunch »