Mar 13, 2019
Craig is on the WGAN Morning News with Ken and Matt. Joe Reagan sitting in for Ken this morning. They talked about the new cameras that can spot a shoplifter even before the steal something. They also discussed autonomous cars and their impact and risks.
These and more tech tips, news, and updates visit - CraigPeterson.com
---
Related Articles:
These Cameras Can Spot Shoplifters Even Before They Steal
Automakers Could Give Police Control Over Your Self-Driving Car
Self-Driving Cars
Might Kill Auto Insurance As We Know It
---
Transcript:
Below is a rush transcript of this segment, it might contain errors.
Airing date: 03/13/2019
Artificial Intelligence Our Safety And Privacy
Craig Peterson 0:00
Hey, good morning everybody. Craig Peterson here on this morning
with WGAN. Ken was out. So we had Joe Reagan in this morning along
with Matt and we had a little bit of chat about a couple of things.
The new AI stuff, Joe had some interesting points today. I enjoyed
having him on. But AI and what does it mean to us when from soup to
nuts here shopping, police, our cars, etc. so here we go with the
guys over at Maine's number one morning show.
Matt Gagnon 0:36
And we're back again. 7:37 on the WGAN Morning News with Ken and
Matt. Joe Reagan is in for Ken today and Craig Peterson is on the
line with us as he would be typically on 7:38 on a Wednesday. So
Craig, are you this morning?
Craig 0:50
Hey, I'm doing good, Matt. What? 737 once you just leave it alone.
Come on, guys.
Matt 0:55
Sorry. I know. A little early this time. So anyway, let's start
with our technology topics. Of course. As always, sir, I think the
first one I'd like to deal with frankly, is is cameras, right.
Spotting shoplifters before they even steal. So basically, we're
talking about what Minority Report here they get, like future
future crimes, pre cognition. The cameras are basically I assumed,
you know, keeping tabs on people and can predict when they're gonna
steal what's happening here?
Craig 1:26
Yeah, this is part of an overall trend that we've been seeing over
the last year or two. We have had in London, of course, we know
right, it's the most surveilled city in the world. They are
tracking you everywhere you go. They're using facial recognition.
And they are using it also to listen listen for bullets been fired,
we have that New York City as well. They know instantly when a gun
is fired, where was fired, we have now software that's being used
by many every major city in the Western world now, that is
predicting where crimes are going to be taking place. And in all of
these cases, it's kind of ok, still, right? Because basically, if
there's an area of the tide crime, you want the police there,
right. And it's kind of a normal thing. Now, we also have
California doing yet another wacky thing, which is instead of
having to go in front of a some form of a magistrate or judge to
get bail, and then you post bail, and you can get out assuming
you're not a big risk to the community or flight risk. California
has decided that they want to get rid of bail bondsmen entirely.
And they just want a computer program to decide who gets to get out
on on basically on bail. And it's it's really concerning now, as we
give more and more power to computer systems which just aren't
fallible, if he wants to any Wi Fi show, right. You know, that's
true.
Matt 3:07
Bad idea.
Joe Reagan 3:07
And then great. You know, one thing that's come up a couple times
is that when you talk about how these computer algorithms are doing
this, so we talked about predictive analysis for shoplifters. A lot
of that is based on correlation. And so therefore, one of the
critiques of the systems is that it will it actually is
discriminatory against usually racial minorities that might be at a
statistically more, you know, our might be punished by criminal by
crime more often. And so, it's, it's, it seems like it almost turns
itself into a, I guess, a self licking ice cream cone, where it's
just, you're just making it worse for people. And it's not actually
doing predictive analysis in terms of someone's intentions, but
actually trying to make judgments based off of past
events.
Matt 3:51
I would like to see a self licking ice cream cone. I would be
quite...
Craig 3:56
Joe, you brought up a really good point. Do you remember Microsoft
came out with this little Twitter bot that they had designed to use
machine learning
Joe 4:09
How can I forget? Like yesterday. Yeah.
Craig 4:11
Yeah, you remember this thing? And what it did was.
Matt 4:13
Yeah. Turned into Skynet in like 10 minutes.
Craig 4:18
Exactly. They had it monitor all these Twitter feeds. And what did
it come back as it came back to as a nasty racist, and it was it
was just crazy. Well, and your point about, you know, predictive
correlation, etc. is a good one, too, because when you start
looking at this in what Matt brought up here was this Japanese
startup called Vaak that takes security camera footage and does
predictions. So you take this and say, well, is this computer
system going to turn into a racist like Microsoft computer system
did because it notices that most of these crimes are committed by
people who have black hoodies on that are covering the faces of
black hoodies. And from a cultural standpoint, that particular type
of dress is, is a minority is, is wearing that type of a dress. So
now all of a sudden, the system that's supposed to be looking for
general body language, general trends is now looking at someone
that walks in that's dressed a certain way and automatically, bam,
it's accusing them of being potential shoplifters. Now, today, what
Bloomberg is reporting is this system is being designed. So that is
alert security and security goes over and asks the shopper, hey, do
you need some help? And that alone is enough for most cases for
them to stop the potential shoplifting that might happen, which is
a big deal guys. We're talking about $34 billion in retail
shrinkage. It's a biggest source, in fact, to shrink. And so it's a
very big deal. But my gosh, where's this all going? We get more and
more of this. And we're not anywhere near artificial intelligence
yet everybody. And what's going to happen everything so far as Matt
pointed out is basically turned into Skynet.
Matt 6:20
Yeah, absolutely. We're talking to Craig Peterson, our tech guru
who joins us at this time to go over what's happening in the world
of technology Okay, so another one that attracted my attention on
our news list here today is the potential future in which cops can
take over my self driving car obviously self driving cars are
inevitable you can be terrified of it if you want to listener who's
listening to this right now, but it's coming and you're gonna have
to get ready for it. So when you finally are forced to drive your
self-driving car when you get pulled over the cops can just like
take over your car man. What's what's what's happening here. The is
this like, be going to be like some sort of stealthy thing bill
into all new cars. You think?
Craig 7:02
Well, that's kind of where we're going right now. And you look at
what happened a couple of weeks ago, in Hawaii, there was a big
conference out there. And they were talking about ethics and the
ethics of artificial intelligence. And one of the things they
brought up was our, you know, home devices like our Alexas, our
Google Home devices, etc. Should have built into them some
artificial intelligence that automatically calls the police if it
detects something that that just might not be a great home
environment, right. As always can figure that out. We already have
these really cool things that reminds me of the Fast and the
Furious, these remote control little little cars that the police
can use. And they drive them underneath the car of a flame suspect.
And they set off a small electromagnetic pulse which disabled the
engine because of course, all these new engines have computers in
them. Have you seen these things?
Matt and Joe 8:09
Yeah, yep.
Craig 8:10
Yeah, little RC cars, and they'll do 80 miles an hour. So in that
case, obviously they're damaging the cars. They just burned out the
computers. And now we have people who are driving a semi autonomous
cars, for instance, the Teslas and there's a lot of Teslas on the
road here in Maine. And they have this kind of semi-auto pilot
thing. And the idea is you can have it kind of take over the
steering wheel. And while you're on the highway, it'll stay in the
lanes. It'll drive down the road. And if the cars in front of you
slow down, it'll slow down. It really kind of makes you a really
good patron of the roads is the idea and Elon Musk has announced
that come next year, he's going to have a software upgrade for the
Teslas that makes them 100% autonomous. But in case now we have a
couple that we know of, we have drivers that have fallen asleep.
And this one case the driver as reported by Bloomberg in December
last year was drunk. Fell sound asleep behind the wheel of his
Tesla. His auto pilot was engaged and the police were chasing them
down the freeway. I'm not sure exactly what they noticed, probably
a sleeping driver. Might be something that's worth noting. And they
will try to get this car to stop it couldn't wake up the the so
called driver, the occupants of the vehicle and so they ended up
having to box the car in. They had, you know, you get four police
officers, the car has to slow down because the cars around it or
slowing down came to a stop. So what do you do. The police are
suggesting and Bloomberg's suggesting that maybe what should happen
here is the the cops should be able to not only have the car pull
over to the side of the road and stop, but they're talking about
rerouting cars, as well as been able to force them to pull over. So
I don't know, Matt, maybe when you're driving that car down the
road, the police will just do that. Let's say Hey, take you right
to the local police station. And because you're under arrest and
all the doors locked.
Matt 10:24
So you're suggesting to me that is that the whole OJ chase thing in
1991 or whatever is the thing of the past. We'll never see that
again.
Craig 10:31
Yeah, exactly. When was the last night you fell drunk on the
road?
Joe 10:36
Well, you know that reason is your question because whether you're
talking about a vehicle that is fully autonomous or partially
autonomous, it raises questions, and this is one of the insurance
industry has dealt with for some time is how do you deal with
damage that's incurred from mechanical issues, whether that be a
computer glitch, or some sort of actual mechanical glitch where the
operator may or may not be 100% responsible for something that's
happens, I get it soon. This is really changing the dynamic of of
how the insurance companies using auto insurance,
Craig 11:06
it will in this case, when you're talking about that, Joe, you were
talking about 90% of the risk, because right now, 90% of the
accidents when they've been been investigated afterwards were
attributed to some form of human error. So if you get rid of that
90%, if it's no longer the human who's driving who is causing the
accident, or at least a large contributor to the accident, where
does that liability go? Because, again, human or you know,
involvement here, right? Human risk from the driver? Well, there
are humans and I'm assuming I, you know, I got a little implication
for what you're saying. I'm assuming that you're also including the
humans that wrote to the software. You know, they don't go after
the guy that designed the switch that failed on the car. And that
caused, you know, I just contributing to an accident. They don't do
that nowadays, right? They might go after the manufacturer, they
might try and get a recall on the car, get that switch replaced,
get that switch repaired. But that guy that misdesigned the switch
doesn't go to jail doesn't face criminal charges. What's the future
going to hold when now we don't even have most of these car
companies making their own software, Apple got rid of their
autonomous vehicle division, at least the guys that were going to
make the cars and they've decided they're going to make the
software so they're not going to make all of the software there's
already dozens of computers in modern cars. There's a lot of
different software, a lot of different companies. Joe, my gosh,
your head's going to explode when you get right down to it and
think about where does a liability go? Who has to get insured? How
long is this stuff going to get tied up in the courts and you know,
Washington DC in state courts and and legislatures. This is this is
the hugest part of this huge new problem we're going to be facing
with autonomous vehicles. All right.
Matt 13:16
Craig Peterson, our tech guru joins us at this time to go over the
world of technology every Wednesday and today is no exception.
Craig appreciate it as always, and we will talk to you again next
week.
Craig 13:26
Hey gentlemen, take care.
---
More stories and tech updates at:
Don't miss an episode from Craig. Subscribe and give us a rating:
Follow me on Twitter for the latest in tech at:
For questions, call or text:
855-385-5553