New Issue: Orbital Catastrophe Ahead? Read Now

How the Testing of Google Duplex Is Coming Along

Join Our Community of Science Lovers!

Since I wrote my column about Google Duplex in Scientific American, Google has cautiously moved ahead with testing this controversial technology. As you may recall, what Duplex does is to make reservations at restaurants and hair salons—by placing a phone call to their human receptionists. It perfectly impersonates a human voice, complete with “um”s, hesitations, and realistic inflections.

When CEO Sundar Pichai first played a recording of a Duplex call at the company’s developer conference, the human receptionist had no idea she was speaking with an AI assistant. That, critics said, was the most objectionable part of the whole thing. Robots should not try to trick humans.

Google Duplex is now in testing in the real world with actual businesses—on a small, slow scale. In this testing phase, the system works only with a hand-picked set of consumers and businesses who both know they’re part of the testing program.


On supporting science journalism

If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.


Furthermore, Duplex now begins the call with full disclosure, along these lines: “Hi, I’m the Google Assistant, calling to make a reservation for a client. This automated call will be recorded. Can I book a table for Tuesday the third?”

(That bit about “will be recorded” plays only in states where notifying your party that you’re recording the call is required by law.)

In other words, there’s no longer any concern that the human receptionist is being “deceived” or “manipulated.”

So why, in that case, does the Google AI voice still include “umms” and “ahs” and human-sounding stumbles and pauses?

Because when Google tried a straight-up digital voice, “it didn’t work,” according to Scott Huffman, VP of engineering for Assistant, in an interview with The Verge. “We got a lot of hangups, we got a lot of incompletion of the task. People didn’t deal well with how unnatural it sounded.”

If the human receptionist goes wildly off-topic or otherwise confuses the AI, the call gets handed off to a team of human operators who take over from there. That backup system kicks in, at the moment, once in every five calls.

Over time, the success rate will improve; Google Assistant will improve; and the world will get used to calls from human-sounding robots. Soon, we won’t object to them any more than we object to voice menus when we call an airline or a corporation—and Google Assistant will continue on its quest to get smarter.

David Pogue is the anchor columnist for Yahoo Tech and host of several NOVA miniseries on PBS.

More by David Pogue
Scientific American Magazine Vol 319 Issue 2This article was published with the title “How the Testing of Google Duplex Is Coming Along” in Scientific American Magazine Vol. 319 No. 2 ()
doi:10.1038/scientificamerican082018-sLDuxEc9Lr6JpRFbPafCf

Subscribe to Support Independent Journalism

Great science journalism requires human expertise, time, effort and creativity. And it costs money. That’s why I and the journalists here at Scientific American hope you’ll join our community.

When you subscribe, you are supporting staff and freelance journalists who are passionate about telling science stories that are true, important and compelling. Our editors and reporters are often experts in their fields, which means they understand the nuances of big discoveries and can untangle the breakthroughs from the hype. With a subscription, you are also supporting rigorous fact-checking to ensure the words we publish are precise and accurate. And you’re supporting original illustrations, graphics and photos that bring you closer to an advanced laboratory, an ice sheet in Antarctica or a space mission in orbit. You’re helping us craft other types of high-quality journalism as well: Our newsletters are carefully written, edited and curated by staffers you have or will come to know and love. Our Science Quickly podcast is based on original reporting, collaboration with editors and scientists and exacting production.

Subscriptions keep this engine running so we can continue to deliver thoughtful, rigorous and independent science journalism to you. In an era of viral misinformation, this work is crucial. If you value what we do, I hope you’ll consider joining us as a subscriber

Thank you,

Jeanna Bryner, Editor in Chief, Scientific American

Subscribe