Friday, May 18, 2018 [Tweets] [Favorites]

Basic Questions About Google Duplex

Dan Primack:

When you call a business, the person picking up the phone almost always identifies the business itself (and sometimes gives their own name as well). But that didn’t happen when the Google assistant called these “real” businesses[…]

John Gruber:

The way the people answered the phone in these recordings was one of the first things that made me suspicious that these examples were either significantly edited or outright fakes. Plus, the salon only asks for a name (and only a first name at that). No phone number, no checking if the client has a request for a certain stylist.

Nick Heer:

Google CEO Sundar Pichai insisted three times that these calls were real, but these discrepancies should be answered. If these calls were edited, even just to remove the business name to limit publicity, Google hasn’t said. Very strange.

Joe Cieplinski:

But if I had to guess: Google made a real phone call, but to someone who had been prepped to follow a very specific script. That way, they were sure to get the responses they wanted. Not so much a complete fake as a contrived circumstance that didn’t demonstrate how this app would behave in the real world.

See also: John Gruber.

Update (2018-06-02): John Gruber:

But Pichai also said “This will be rolling out in the coming weeks as an experiment.” On the one hand, that makes me feel like maybe I am off my rocker for being so skeptical. Why in the world would Pichai say that if they weren’t at a stage in internal testing where Duplex works as the recordings suggest? But on the other hand, if they are that close, why haven’t they invited anyone from the media to see Duplex in action?

[…]

The headlines last week should have been along the lines of “Google Claims Assistant Can Make Human-Sounding Phone Calls”, not “Google Assistant Can Make Human-Sounding Phone Calls”. There’s a difference.

2 Comments

It seems pretty obvious that the calls were edited to remove the business names. The idea that Google is outright lying, and that these calls were staged, sounds like a bit of a conspiracy theory. It seems like it would be a pretty bad idea for Google to just flat-out lie to people about something like that. The calls also don't sound particularly staged, with mumbling answers and the AI and the person at the other end kind of talking past each other.

What does bother me is something else: Google is clearly trying to make the AI sound like a human, with "uhms" and stuff like that added to their responses. This seems kind of unethical to me. I don't know, maybe we have to figure out what exactly is the okay thing to do here, but it seems like a bad idea to make the AI pretend to be human.

This controversy is a bit absurd regarding some of the questions asked. In particular the part about the calls being "recorded without consent".

- You can't blame Google at the same time for not identifying the businesses or the employees and having hypothetically recording the calls without consents.

- You can't question this part and just forget that for instance, some guy in a jean and a black mock turtleneck made his first call with a smartphone to Starbucks to order 4,000 lattes and this was done in front of a big audience and is still available because it's been recorded. And based on this story, Apple did call and record without consent. https://www.fastcompany.com/3006147/because-steve-jobss-first-public-iphone-call-starbucks-still-

Stay up-to-date by subscribing to the Comments RSS Feed for this post.

Leave a Comment