WEBVTT - Google Duplex Takes Robocalls Next Level

0:00:04.720 --> 0:00:07.160
<v Speaker 1>More on that Google phone call that rocked the world.

0:00:07.280 --> 0:00:10.440
<v Speaker 1>I'm Rich Demiro. This is Rich on Tech Daily at

0:00:10.480 --> 0:00:13.000
<v Speaker 1>Google Io. Google made a phone call that got quite

0:00:13.039 --> 0:00:16.599
<v Speaker 1>the reaction. It was Google assistant calling a hair salon

0:00:16.720 --> 0:00:18.440
<v Speaker 1>to make an appointment. Take a listen.

0:00:21.800 --> 0:00:24.880
<v Speaker 2>So happening out here. Hi, I'm calling to book a

0:00:24.920 --> 0:00:27.800
<v Speaker 2>women's haircut for a client. I'm looking for something on

0:00:27.920 --> 0:00:31.600
<v Speaker 2>May third. Sure I give me one second?

0:00:32.720 --> 0:00:33.320
<v Speaker 3>Mm hmm.

0:00:36.080 --> 0:00:39.559
<v Speaker 2>Sure? What time are you looking for a round at

0:00:39.600 --> 0:00:43.200
<v Speaker 2>twelve pm? We do not have a twelve pm available.

0:00:43.440 --> 0:00:46.599
<v Speaker 2>The closest we have to that is a one fifteen.

0:00:47.640 --> 0:00:51.720
<v Speaker 2>Do you have anything between ten am and twelve pm?

0:00:51.760 --> 0:00:54.880
<v Speaker 2>Depending on what service she would like? What service is

0:00:54.880 --> 0:00:59.240
<v Speaker 2>she looking for? Just a woman's haircut for now? Okay,

0:00:59.280 --> 0:01:03.240
<v Speaker 2>we have a ton of can I am a fine? Okay?

0:01:03.280 --> 0:01:08.480
<v Speaker 2>What's your birth name? The first name is Lisa. Okay, perfect,

0:01:08.520 --> 0:01:13.559
<v Speaker 2>So I will see Lisa at ten o'clock on May third. Okay, great, thanks, great,

0:01:13.720 --> 0:01:14.600
<v Speaker 2>have a great day. Back.

0:01:16.080 --> 0:01:19.880
<v Speaker 1>Yes, that's artificial intelligence at work, a computer calling a

0:01:19.920 --> 0:01:22.720
<v Speaker 1>real person and doing all the back and forth to

0:01:22.760 --> 0:01:26.560
<v Speaker 1>set up an appointment. That are even the hums and mms.

0:01:27.000 --> 0:01:30.440
<v Speaker 1>The project is called Google Duplex. My post on Twitter

0:01:30.480 --> 0:01:35.240
<v Speaker 1>got lots of reactions. Luis Claudio Costa said, Siri, You're fired.

0:01:35.680 --> 0:01:38.760
<v Speaker 1>Dylan Robertson said it's so realistic that I thought it

0:01:38.800 --> 0:01:42.320
<v Speaker 1>was some kind of prank video. Mind blown, and Gregory

0:01:42.400 --> 0:01:44.560
<v Speaker 1>Dyke said last time I was impressed that much was

0:01:44.600 --> 0:01:48.120
<v Speaker 1>when the first iPhone was presented. Wow. Now you can

0:01:48.200 --> 0:01:50.120
<v Speaker 1>draw your own conclusions whether you think this is the

0:01:50.160 --> 0:01:52.920
<v Speaker 1>best technology in the world or kind of scary. But

0:01:53.000 --> 0:01:55.559
<v Speaker 1>after the io QINOT I got to talk to Rishi Chandra,

0:01:55.600 --> 0:01:58.280
<v Speaker 1>who obviously is very excited about this. He's the vice

0:01:58.320 --> 0:02:01.720
<v Speaker 1>president of Product Management and general manager of Home Products.

0:02:01.920 --> 0:02:05.440
<v Speaker 3>Again, it's about connecting your goals and intent with the

0:02:05.480 --> 0:02:07.520
<v Speaker 3>businesses that actually can fill those intents. So when I

0:02:07.560 --> 0:02:09.800
<v Speaker 3>want to order flowers or book a restaurant or what

0:02:09.880 --> 0:02:12.160
<v Speaker 3>have you, we want to make it as easy as possible,

0:02:12.480 --> 0:02:15.959
<v Speaker 3>and today it's not that easy. And so by connecting

0:02:16.040 --> 0:02:17.799
<v Speaker 3>kind of consumers with businesses, we think we can help

0:02:17.800 --> 0:02:21.080
<v Speaker 3>both people. Right, businesses have more customers coming in and

0:02:21.120 --> 0:02:25.239
<v Speaker 3>consumers have easier access to all the great available technology

0:02:25.280 --> 0:02:26.280
<v Speaker 3>or services around them.

0:02:26.600 --> 0:02:29.760
<v Speaker 1>Rischie explained how Duplex can help with all those businesses

0:02:29.800 --> 0:02:32.360
<v Speaker 1>out there that aren't on Open Table or a similar

0:02:32.400 --> 0:02:33.639
<v Speaker 1>online booking platform.

0:02:33.720 --> 0:02:35.520
<v Speaker 3>I think one of the big insights we found is

0:02:35.560 --> 0:02:39.400
<v Speaker 3>that not every business today is digitally connected right. In fact,

0:02:39.400 --> 0:02:41.440
<v Speaker 3>the vast majority are still not digitally connected, and so

0:02:41.440 --> 0:02:42.960
<v Speaker 3>how do you reach them? How do we help you

0:02:43.040 --> 0:02:47.000
<v Speaker 3>assist you in ways of getting haircut appointments or booking

0:02:47.000 --> 0:02:49.840
<v Speaker 3>a restaurant reservation. And so the insight was, well, look,

0:02:49.960 --> 0:02:52.040
<v Speaker 3>everyone has phones today, and how do we make it

0:02:52.080 --> 0:02:54.320
<v Speaker 3>so that ORAI can help you kind of take advantage

0:02:54.320 --> 0:02:57.040
<v Speaker 3>of assisting you using the phone. But it's still really

0:02:57.320 --> 0:02:59.440
<v Speaker 3>early technology. But some of the examples you saw is

0:02:59.480 --> 0:03:02.520
<v Speaker 3>how quickly it's advanced. The ability for the assistant to

0:03:02.560 --> 0:03:05.160
<v Speaker 3>really react and be able to have a natural conversation

0:03:05.240 --> 0:03:07.000
<v Speaker 3>so on is something a few years ago we never

0:03:07.040 --> 0:03:09.680
<v Speaker 3>thought possible. But we're at that forefront now and so

0:03:10.040 --> 0:03:11.880
<v Speaker 3>we're excited to experiment with this. We're going to launch

0:03:11.919 --> 0:03:14.200
<v Speaker 3>it as an experiment next few weeks and we'll see

0:03:14.200 --> 0:03:16.280
<v Speaker 3>how it goes. But I think we believe we are

0:03:16.320 --> 0:03:18.400
<v Speaker 3>actually the beginning of something much bigger where we can

0:03:18.440 --> 0:03:20.560
<v Speaker 3>assist you in so many different ways that we could

0:03:20.560 --> 0:03:21.000
<v Speaker 3>in before.

0:03:21.360 --> 0:03:23.480
<v Speaker 1>One of the big questions, will the person at the

0:03:23.560 --> 0:03:26.680
<v Speaker 1>business who's answering the phone know that they're talking to

0:03:26.720 --> 0:03:27.280
<v Speaker 1>a computer.

0:03:27.560 --> 0:03:28.560
<v Speaker 3>One of the things we want to do is make

0:03:28.560 --> 0:03:31.040
<v Speaker 3>sure we're being very transparent obviously with the businesses, and

0:03:31.120 --> 0:03:33.400
<v Speaker 3>so we're actually experiment with the different ways of saying

0:03:33.400 --> 0:03:35.440
<v Speaker 3>that you'll see actually in the keynote they said, Hey,

0:03:35.480 --> 0:03:38.280
<v Speaker 3>I'm calling on behalf of this person. I'm a client

0:03:38.440 --> 0:03:41.120
<v Speaker 3>of behalf that person. That's just one example. There's multiple

0:03:41.160 --> 0:03:43.240
<v Speaker 3>examples of things we're actually testing out and trying. This

0:03:43.320 --> 0:03:45.160
<v Speaker 3>is a whole purpose of the experiment to make sure

0:03:45.200 --> 0:03:48.240
<v Speaker 3>that everyone feels that we are being fully transparent and

0:03:48.240 --> 0:03:50.600
<v Speaker 3>that even the restaurants feel it and they understand what's

0:03:50.600 --> 0:03:53.000
<v Speaker 3>actually happening. And restaurants of course will also be able

0:03:53.040 --> 0:03:55.440
<v Speaker 3>to opt out if they're not uncomfortable technology for whatever reason.

0:03:55.760 --> 0:03:59.240
<v Speaker 1>Google also announced six new voices for Google Insistent, including

0:03:59.280 --> 0:04:00.280
<v Speaker 1>a celebrity voice.

0:04:00.560 --> 0:04:00.720
<v Speaker 2>Now.

0:04:00.720 --> 0:04:03.360
<v Speaker 1>The reason for this it's easier than ever for their

0:04:03.400 --> 0:04:05.200
<v Speaker 1>computers to synthesize voices.

0:04:05.440 --> 0:04:07.360
<v Speaker 3>Generally, what we've done is you need a seed, so

0:04:07.440 --> 0:04:11.280
<v Speaker 3>you have kind of a set of voices that are scripted,

0:04:11.680 --> 0:04:13.920
<v Speaker 3>and then based off that we can start programming and

0:04:13.960 --> 0:04:16.400
<v Speaker 3>so they can almost say anything, which is great because

0:04:16.400 --> 0:04:18.600
<v Speaker 3>obviously the assistant can respond in so many different ways

0:04:18.680 --> 0:04:20.960
<v Speaker 3>or to be Google Search responses or Google Maps responses.

0:04:21.520 --> 0:04:23.520
<v Speaker 3>And what we've done is we've trained a bunch of

0:04:23.640 --> 0:04:25.919
<v Speaker 3>neural nets that actually allow us to build custom voices

0:04:25.920 --> 0:04:28.159
<v Speaker 3>in a pretty seamless way, and we can actually have

0:04:28.200 --> 0:04:31.400
<v Speaker 3>a wide diversity. Right Why should assistant only have one voice?

0:04:31.560 --> 0:04:33.479
<v Speaker 3>And one of the most exciting things we are talking

0:04:33.520 --> 0:04:35.560
<v Speaker 3>about today is that John Legend actually is gonna be

0:04:35.600 --> 0:04:37.880
<v Speaker 3>one of those seeded voices, and so you're gonna have

0:04:37.960 --> 0:04:40.200
<v Speaker 3>John Legend actually talk to you. We're actually pretty excited

0:04:40.200 --> 0:04:41.960
<v Speaker 3>to get that out the door. So the six voices

0:04:41.960 --> 0:04:43.600
<v Speaker 3>are going to launch today, and then John Legend will

0:04:43.640 --> 0:04:44.760
<v Speaker 3>actually launch later this year.

0:04:45.040 --> 0:04:46.640
<v Speaker 1>And by the way, if you want to change your

0:04:46.640 --> 0:04:49.680
<v Speaker 1>Google Assistant Voice right now, just go into Google Assistant,

0:04:49.800 --> 0:04:53.839
<v Speaker 1>then Preferences, then Assistant Voice, and you've got eight voices

0:04:53.839 --> 0:04:57.120
<v Speaker 1>to choose from. Have fun. Speaking of Assistant, Google also

0:04:57.160 --> 0:05:00.520
<v Speaker 1>announced a feature called pretty Please, which is helpful for kids.

0:05:00.760 --> 0:05:03.040
<v Speaker 3>You know, it's been one of the top feature requests.

0:05:03.080 --> 0:05:06.040
<v Speaker 3>I think even my wife notices, you know when one

0:05:06.080 --> 0:05:07.800
<v Speaker 3>of the things that kids love about the product is

0:05:07.800 --> 0:05:09.760
<v Speaker 3>that they for the first time they're in charge. They

0:05:09.800 --> 0:05:13.279
<v Speaker 3>get to tell a device what to do. But of

0:05:13.320 --> 0:05:14.800
<v Speaker 3>course one of the things we want to train you

0:05:15.000 --> 0:05:16.760
<v Speaker 3>kids on is how to do it in an appropriate way.

0:05:17.120 --> 0:05:19.400
<v Speaker 3>And so the pretty Preeace feature is a way that

0:05:19.440 --> 0:05:22.520
<v Speaker 3>we can get positive reinforcement to kids who actually say

0:05:22.520 --> 0:05:25.040
<v Speaker 3>please or thank you when they're actually interacting with the assistant.

0:05:25.120 --> 0:05:27.360
<v Speaker 3>We think it's an another mechanism that we can help

0:05:27.400 --> 0:05:30.440
<v Speaker 3>make sure that we're having technology really work in people's lives,

0:05:30.760 --> 0:05:33.080
<v Speaker 3>and that enhances kind of the kids' experience with the product.

0:05:33.560 --> 0:05:36.280
<v Speaker 1>There you have it, robocalls two point zero, but this

0:05:36.440 --> 0:05:39.200
<v Speaker 1>time they might actually be kind of useful. Thanks so

0:05:39.279 --> 0:05:42.200
<v Speaker 1>much for listening to the podcast and leaving those ratings

0:05:42.240 --> 0:05:46.400
<v Speaker 1>and reviews in the Apple podcast sapp like Airwaves seventy five,

0:05:46.480 --> 0:05:49.520
<v Speaker 1>who said this is an awesome podcast that keeps me

0:05:49.720 --> 0:05:52.240
<v Speaker 1>up to date in the tech world. Rich Demiro does

0:05:52.279 --> 0:05:56.240
<v Speaker 1>an exceptional job bringing that information to the listeners. Highly recommended.

0:05:56.440 --> 0:05:58.880
<v Speaker 1>Thanks so much, Airwaves seventy five. And if you leave

0:05:58.920 --> 0:06:01.080
<v Speaker 1>a review, I might just read it right here on

0:06:01.160 --> 0:06:03.440
<v Speaker 1>the show. I'm Rich Damiro links to everything I talk

0:06:03.480 --> 0:06:07.080
<v Speaker 1>about here on my website. Just go to richontech dot tv.

0:06:07.480 --> 0:06:09.000
<v Speaker 1>I'll talk to you real soon.