Name: How A.I. is evolving in a post-pandemic world
Uploaded: 2021-11-09
Duration: 4 min 21 s
Description: Leaders at Amazon and Spotify tell us how demand for customized experiences is a driving force.

Leaders at Amazon and Spotify tell us how demand for customized experiences is a driving force.

How A.I. is evolving in a post-pandemic world

Conferences

we both talked about trying to make you know this recommendation experience more human like I mean you know in the case of Spotify I guess one of things you guys have done is actually kind of keep humans in the loop to some extent. Um I know you have this thing you refer to as alga tutorial. I wonder if you could talk a little bit about what that is the odd and explain to people how that works. Yeah, of course. Um so we tried to coin some words. So algo tutorial is basically the combination of algorithmic and editorial and it is really the principle that the whole is greater than the sum of the parts. If you take all the power of our technology and personalization and you bring in all the expertise and the passion of hundreds of editors that we have around the world who are experts in music. So when you put those two things together you will have a better experience than if you took them separately in the case of advertorial. One very big success that we've had is in creating playlists that are alga tutorial. And so what that means just as an example is you take a playlist called songs to sing in the shower or songs to sing in the car. There are like these hypotheses That you know are attributes of songs that experts are able to pick. But like let's say the expert picks the 500 songs that are the quintessential songs you want to sing on a road trip, then they have done an amazing job picking and defining the corpus but but if you don't know those songs, it is useless for you. So this is where we come in with our personalization technology and we refine the selection and we make it personalized um playlist for you powered by technology but informed by the expertise of users of editors. So this is just an example of advertorial and we tend to think that when you bring these things together the experience gets better. So the thing we are doing this year, he's not only putting humans in the loop with editors, but we're bringing humans in the loop with users. So we used to have either made for you playlists and experiences or made by you. So made for you is when we build the daily mix for you. Made by you is when you create your own playlists. What we are doing this here is bringing these two things together. So we had to launch earlier a few months ago where it enhances your playlist, you can opt into it. But if you hit enhance on your own playlist, we will bring recommendations in your playlists to help you build it that are based on the ones you have already selected. So this is basically the user and the machine learning. Hand in hand building the experience. So we tend to try to do these things and Rohit and Alexa. How are you trying to make elections seem more human, like more conversational. Can you talk to us about where the cutting edge is there? Yeah, I think first our vision is for Alexa to be that ambient artificial intelligence for you, which is there when you need it and recedes into the background when you don't for that. I think we are the only way to solve that problem is more with the more generalized intelligence methodologies where Alexa learns fast about many different tasks. It already has 100 30,000 skills. Imagine the way that's growing our expectations of our customers is that Alexa should be able to do anything possible for that. The only way to do that is more multi task intelligence. Then the second aspect of it is it needs to be more self learning when it engages with the environment and understands you and gets better every time. And then it needs to anticipate your needs. Uh think about if you left your garage door open and Alexa should be notifying you. That means it needs to be more self aware and in particular aware of your ambience state to help you with that. But lastly, I think when you have a relationship with the customer where they are really trying to get the best utility from the ai is for it to be self. So this is where we believe that the way to make this intelligence go forward is for the customers to be able to directly teach Alexa, which means Alexa should become more self serve, not just by developers all around the world, but also by the end user, where you can say, uh, let me know when the dog is barking or when when you saw your kids playing Xbox and you want Alexa to let you know that they're playing too much video games. It's a two problem in March household, so that's the kind of intelligence that we are aspiring to where the eye can be truly useful for everyone.