- Curvy Wife Guy makes the most Curvy Wife Guy pregnancy announcement 10 Months Ago
- ‘A Fortunate Man’ is a gorgeous, shallow adaptation of a classic novel 10 Months Ago
- Obama’s ‘Easter worshippers’ tweet upsets conservatives 10 Months Ago
- Trump responds to impeachment calls: ‘You can’t’ 10 Months Ago
- Desperate YouTuber goes on ‘holy pilgrimage’ to meet PewDiePie 10 Months Ago
- Influencer faces 20 years for hiring cousin to threaten man into giving up domain name Today 10:17 AM
- ‘Avengers: Endgame’ isn’t actually the end of Marvel’s Phase 3 Today 10:12 AM
- ‘Star Trek: Discovery’ just obliterated every stupid argument about its place in canon Today 9:24 AM
- ‘Game of Thrones’ brings us lots of horny Tormund memes Today 8:46 AM
- 12 adult toys sex bloggers swear by Today 8:45 AM
- Howard Schultz mocked for ‘majority of Americans are Americans’ ad Today 8:35 AM
- The ultimate guide to strap-ons Today 8:00 AM
- Why ‘Ant-Man and the Wasp’ is my favorite Marvel movie Today 7:59 AM
- The quick and dirty guide to buying your first vibrator Today 7:45 AM
- How to watch ‘Live Rescue’ for free Today 7:00 AM
How Google is making Duplex, its AI calling system, less creepy
When Google debuted its Duplex artificial intelligence, an automated phone calling system for scheduling appointments, the public reaction was pretty unanimous. Yes, this is incredibly cool—and impressively realistic—but also intensely creepy. Now we know how Google is fighting that initial impression and ensuring its Duplex AI doesn’t violate human privacy expectations.
To better explain how Duplex works, Google gave a hands-on demonstration with the software, the Verge reports.
At the start of every call, Duplex will announce itself with a message. In one example, the Verge said it went along the lines of, “Hi, I’m calling to make a reservation. I’m Google’s automated booking service, so I’ll record the call. Uh, can I book a table for Sunday the first?”
There are different versions of this message—Google may be testing which versions make the human on the end of the line feel the most comfortable—but each of these disclaimers was followed by a human-like “uh” or “um.” (The reason for that human-like tic is interesting: Apparently, call answerers did not respond well when the voice was more robotic. The pause also gives answerers a moment to respond and say, “No thanks, I don’t want to be recorded.” In this case, Duplex hangs up, and a human operator calls back instead.)
Google is only rolling out Duplex to a small number of “trusted” businesses initially. That is, these establishments and their employees will know that they’ll be receiving automated Duplex calls, which means it won’t come as a surprise or a deception. Businesses will also be able to opt out of Duplex altogether.
At first, Google is limiting Duplex to inquiries about business’ holiday hours, then will advance to restaurant reservations, and then will tackle hair appointments.
While the AI can get confused by some human interactions and decide to end the conversation, overall, it’s good at its job. Per the Verge:
It handled a bunch of interruptions, out of order questions, and even weird discursive statements pretty well. When a human sounded confused or flustered, Duplex took a tone that was almost apologetic. It really seems to be designed to be a super considerate and non-confrontational customer on the phone.
Google introduced the Duplex AI at its annual developer conference in May. Shortly after that onstage demo, Google was forced to backtrack due to reactions from the press and the public, which questioned whether the service was ethically sound. Google said later assured disclosure would be built into the product, which we’ve now seen with this latest demo.
All-in-all, Duplex looks to be a promising way to handle what can be a time-consuming task—but as for whether it’s more efficient than making an appointment online yourself, we’ll have to wait and see.
H/T the Verge
Christina Bonnington is a tech reporter who specializes in consumer gadgets, apps, and the trends shaping the technology industry. Her work has also appeared in Gizmodo, Wired, Refinery29, Slate, Bicycling, and Outside Magazine. She is based in the San Francisco Bay Area and has a background in electrical engineering.