www. O S N E W S .com
News Features Interviews
BlogContact Editorials
.
Google Call Screen: a robot that will answer spam calls
By Thom Holwerda on 2018-10-09 23:46:58

Not everything got leaked before Google's event today. One surprise announcement that wowed was Call Screen, a new feature that lets the Google Assistant answer your incoming calls and politely ask what the caller wants. A real-time transcript will appear on your screen, allowing you to decide whether or not you want to pick up.

When your Pixel rings, a "Screen call" button shows up alongside the usual controls. Tapping it will prompt the Google Assistant to tell your caller that the call is being screened and ask what it's about. Their explanation is transcribed on your screen, and you have options to mark the call as spam or tell the caller you'll get back to them, among others.

This is an amazing feature that will save a lot of people a lot of frustration. I want this feature on my phone now.

On a related note, Google Duplex, the feature whereby the Google Assistant will call restaurants and such on your behalf, will be rolled out to Pixel phones next month.

 Email a friend - Printer friendly - Related stories
.
Post a new comment
Read Comments: 1-10 -- 11-20 -- 21-27
.
RE: Google's AI isn't necessary
By Alfman on 2018-10-11 14:14:51
M.Onty,

> I like the idea. But it doesn't need AI, not really. Its just being used because Google spent a lot of money developing some AI, now they need to find a use for it. That's fine, but its not needed.

Ever been fooled by one of your smart-arse friends who has an answer-machine in their voice that says something like, "Oh hi... Yep. Uh huh. Yep... Oh actually I can't get the the phone right now"?


That's exactly what's going on here... users record a customized "script", and it wastes the telemarketer's time. Despite being humorous, it's remarkably effective.

https://www.youtube.com/watch?v=T...
https://www.youtube.com/watch?v=_...

After wasting the caller's time, one of the recordings ends with "Ok you know what? I have to go gargle some eggs, so I have to cut this conversation short but thanks for listening."
Not a bad way of dealing with pests. Main problem I see with it is that one doesn't necessarily know whether a call is legit or not before answering.



I've run across this channel before. It's a guy who has access to a bunch of phone numbers, so he decided to have his robot answer the calls and run through some quasi intelligent scripts. His bot dials digits to get to an operator.
https://www.youtube.com/channel/U...

It shows that you don't need a very sophisticated AI to combat telemarketing.
Permalink - Score: 4
.
RE: But in what accent?
By Kochise on 2018-10-11 15:25:27
https://www.youtube.com/watch?v=-...
Permalink - Score: 2
.
RE[3]: Defeats the point?
By zima on 2018-10-11 17:23:30
Well, then those will be the numbers which you will ignore / to which you just won't call back...
Permalink - Score: 2
.
RE[2]: Won't help
By zima on 2018-10-11 17:25:08
But at some point those calls could make your number too often unreachable by other people?... :P
Permalink - Score: 3
.
RE[2]: Google's AI isn't necessary
By M.Onty on 2018-10-11 18:18:32
Thanks for sharing that!
Permalink - Score: 2
.
RE[4]: Defeats the point?
By computrius on 2018-10-11 18:38:58
Correct. So, when compared to voicemail, it changes nothing.
Permalink - Score: 1
.
RE[4]: Defeats the point?
By Alfman on 2018-10-11 18:46:40
zima,

> Well, then those will be the numbers which you will ignore / to which you just won't call back...

Ignoring the problem hasn't made the problem go away. Blocking or not answering calls based on caller id is an obvious strategy, however scammers have long since evolved beyond it. Now I'd say nearly all of these calls are from random "local" numbers to make them appear legit. These are real local numbers in our area, the scammer just impersonates local numbers randomly. Blocking these numbers not only does nothing at all to stop the calls, but it potentially blocks calls from legitimate people in the future. I've gotten callbacks from annoyed people who say they received unsolicited calls from me, and I know that means that my number was used fraudulently by a scammer.


Here's an interesting twist, I moved from a different state over a decade ago, but I kept my phone number. All the scammers use that area code to impersonate caller id, yet nearly all legitimate calls are from the new york area with completely different area codes. In others words, because my area code is "wrong", I could actually identify most scam calls when I see it's from my own area code, haha.


Anyways, the fundamental problem underlying all of this is that caller-id doesn't reflect the true caller. The telephone companies do receive accurate information for billing purposes (the ANI number). But this number gets stripped off for the subscriber, we are only allowed to see the CID information.

http://www.tech-faq.com/ani-auto...
https://en.wikipedia.org/wiki/Aut...

Alas this is the fundamental reason that device based blocking technology (apps or PBX) does not effectively block scammers. The true originators could be blocked, but it needs to be based on ANI, which the telephone companies don't share. There's one general exception to this, and that's with 800 numbers. The callee has the right to know who's calling because they're paying for the call. Blocking/faking caller-id doesn't work with 800 numbers.


Most calls I get are unsolicited. While we theoretically have the right to sue unsolicited commercial callers for ignoring the do-not-call list, the CID mechanism is basically useless in identifying the real callers. Even reporting scam calls to the FTC is useless when the CID is fake. The technical solution is clear, identify callers using the real information in the ANI that was used to complete the call. The call cannot succeed without it (the telco wouldn't get paid. They have a large incentive to stop that quick). Unfortunately residential numbers don't receive this information and the telephone companies have done squat for blocking scam calls using the ANI information on their end.

So I don't consider it a technical problem so much as a bureaucratic one. Solutions that use CID are all going to fall short. Screening calls with AI may help, but IMHO it's a very indirect way to go about solving the problem.

Edited 2018-10-11 18:57 UTC
Permalink - Score: 2

Read Comments 1-10 -- 11-20 -- 21-27

Post a new comment
Username

Password

Title

Your comment

If you do not have an account, please use a desktop browser to create one.
LEAVE SPACES around URLs to autoparse. No more than 8,000 characters are allowed. The only HTML/UBB tags allowed are bold & italics.
Submission of a comment on OSNews implies that you have acknowledged and fully agreed with THESE TERMS.
.
News Features Interviews
BlogContact Editorials
.
WAP site - RSS feed
© OSNews LLC 1997-2007. All Rights Reserved.
The readers' comments are owned and a responsibility of whoever posted them.
Prefer the desktop version of OSNews?