From NPR News, this is ALL THINGS CONSIDERED. I'm Melissa Block.


I'm Robert Siegel and it's time now for All Tech Considered. Advertisers want to hear what you have to say. As NPR's Steve Henn reports, Madison Avenue and Silicon Valley are about to roll out new kinds of ads you can actually converse with. Marketers are hoping to leverage the power of voice and the kinds of technologies that power Apple's Siri to start selling us all sorts of things.

STEVE HENN, BYLINE: Imagine for a second what it would be like if you could talk to your radio, and your radio would actually listen. To get an idea of what that might be like, last week I downloaded an app called Talk to Esquire, from the magazine, and when I opened it, the app asked me a question.

DAVID WONDRICH: What's your favorite type of liquor? Vodka....

HENN: OK. So that's a little forward, but it's Esquire so. Actually, I'm more of a beer drinker.

WONDRICH: Well, like they say, all that beer and no whiskey, I think you are making a mistake.

HENN: Before I knew it, I found myself engrossed in a chat about booze. I was talking to recordings of a real person, Esquire's DavidWondrich, the drinks columnist. He, or I guess really the app, was listening to what I said and it was answering me back. I think I'd like something with a flowery umbrella.

WONDRICH: Good to hear. So I'd like to mix you a classic cocktail...

HENN: This went on for more than four minutes, and eventually my cyber David Wondrich showed me how to mix a mezcal old fashioned.

WONDRICH: Stir that.

HENN: Professor Clifford Nass of Stanford says this kind of synthetic conversation is possible because of advances in voice and language recognition.

CLIFFORD NASS: It definitely has gotten much better.

HENN: It's no longer just a party trick, but it still has limits.

It's still not super accurate, nowhere near the accuracy of even a 2-year-old child.

Nass says an app like Esquire's can only work well when it tightly limits what the conversation is about. Without context, apps like these are lost. Still, many marketers see an opportunity. Some of the biggest technology companies in the country, from Facebook and Google to Microsoft and Amazon, are trying to figure how to best monetize mobile advertizing.

MIKE MCSHERRY: Mobile banner ads just have not been super successful.

HENN: Mike McSherry is a vice president at Nuance, a company that specializes in voice-recognition technology.

MCSHERRY: Voice is providing this window to the world of information, and it can be far faster than typing or navigating small screens, which has been typically the challenge of small mobile devices to date.

HENN: EMarketer, the digital forecasting company, estimates mobile ads could be a $37 billion business in the next few years. Nuance wants a piece of that. This month, it released a mobile ad platform that will let advertisers create ads we can talk to. I played one of their demos, called Valentine's Dog House, for Clifford Nass.

UNIDENTIFIED WOMAN: Who are you in trouble with?

NASS: My wife.

UNIDENTIFIED WOMAN: And how mad is she?

HENN: Folks at Nuance say ads like these could work on digital radio stations or launch after a click. And even though Professor Nass was sitting at his desk with a lollipop in his mouth, mumbling, this ad nailed it. Nass smiled to himself. He chuckled. He engaged with it. Nass says ads like these, at least when they work, have the potential to be incredibly powerful because they interact with us in a profoundly human way.

NASS: The human brain is built for speech, so anything that sounds like a voice, our brains just light up and we get an enormous range of social and other responses.

HENN: A voice, even a synthetic one, is packed with information our brains are programmed to decode, like gender or age. We can't even help ourselves. So when we talk to a computer and it answers back, Nass says most of us end up acting as if we're dealing with a real person. Social norms kick in seemingly automatically.

NASS: Our brains are built to treat these conversations with computer-based voices to an incredible degree like we were having with actual people, including flattery, flirtation and all the rest. We will see all of those same responses.

HENN: Nass' own research has found that people like computers better if those computers flatter them. He found that we're polite to the machines we talk to, and we identify with computerized voices that sound kind of like our own. We're even more likely to buy stuff from them. Advertisers are hoping they will be able to use all of this, but Nass warns there are big risks in these kinds of interactive ads as well.

NASS: When they work well, they're fantastic; when they work poorly, they're really insulting and disturbing.

HENN: Nass says if a computer sounds like a person, but then interrupts you and acts like a jerk, real human beings are going to be much more likely to hold a grudge. Steve Henn, NPR News, Silicon Valley.

Copyright © 2013 NPR. All rights reserved. Visit our website terms of use and permissions pages at for further information.

NPR transcripts are created on a rush deadline by a contractor for NPR, and accuracy and availability may vary. This text may not be in its final form and may be updated or revised in the future. Please be aware that the authoritative record of NPR’s programming is the audio.



Please keep your community civil. All comments must follow the Community rules and terms of use, and will be moderated prior to posting. NPR reserves the right to use the comments we receive, in whole or in part, and to use the commenter's name and location, in any medium. See also the Terms of Use, Privacy Policy and Community FAQ.

NPR thanks our sponsors

Become an NPR sponsor

Support comes from