Home
Đăng nhập
Đăng ký
Loading...
Luyện nghe
Luyện nghe
/
Video
/
TED Talk
/
Can a Camera on Every Corner Make Us Safer? | Garrett Langley | TED
Can a Camera on Every Corner Make Us Safer? | Garrett Langley | TED
Chọn chế độ học:
Xem phụ đề
Chọn từ
Viết lại từ
Highlight:
3000 Oxford Words
4000 IELTS Words
5000 Oxford Words
3000 Common Words
1000 TOEIC Words
5000 TOEFL Words
Phụ đề (420)
0:04
I'm going to start with a story um
0:07
speaking to a grandmother in rural
0:10
Tennessee and she was taking care of her
0:12
11-year-old granddaughter.
0:15
So, she wakes up, she walks into her
0:17
room.
0:19
The bed's empty. She looks around.
0:22
There's no note. There's nothing. If
0:24
your heart's not racing right now,
0:26
should be. She calls 911.
0:29
detective arrives. No sign of force.
0:33
There's nothing except one thing that he
0:35
remembered. There's a flock camera just
0:37
down the street. So, he pulls out his
0:40
phone and runs a search for the night
0:43
and finds one car.
0:45
He runs that tag and it's his worst
0:48
nightmare. It's a registered sex
0:51
offender. So, he logs into Flock and and
0:53
he's looking for this car and he sees
0:54
the car is traveling down I75. If you're
0:57
not familiar with Tennessee and Georgia,
0:58
that's the connecting state. It's the
1:01
highway. The person's on the highway.
1:03
The detective doesn't know what to do,
1:05
so he does whatever hero does is he
1:07
moves into action. He jumps in his car
1:10
and races down I75.
1:12
As he approaches the interstate line, he
1:15
sees the vehicle.
1:18
Violent struggles pursues and at the
1:21
end, he hears crying.
1:23
And the girl's in the back. She's bound,
1:27
but she's alive.
1:29
They went on to go search the suspect's
1:32
house,
1:33
and the nightmare got scarier.
1:36
Everything one would need to not only
1:38
assault, but dispose of the body was
1:40
there.
1:42
Now, that girl's alive today because a
1:43
detective had two things. He had a
1:46
license plate and he had direction to
1:48
travel. This, for me, is one of
1:51
thousands of stories I sadly hear about
1:53
every day in America. I started Flock 9
1:56
years ago because of this exact type of
1:57
problem. See, I live in in a city called
1:59
Atlanta, and I was frustrated. Crime
2:03
felt too easy to get away with. It
2:05
didn't make sense to me that you could
2:07
steal someone's car and have an 80%
2:10
chance of getting away with it. Or you
2:12
could kill someone and have a 50% chance
2:13
of getting away with it. So, I kept
2:15
pushing. I kept pushing and eventually I
2:17
found a major Atlanta Police Department
2:19
who told me, "Mr. Langley, it's not
2:21
apathy. It's not effort, it's evidence.
2:26
If all we had was a license plate, we
2:28
could catch these people. So, I I did
2:30
what engineers did. I built a license
2:33
plate reading camera. It sounds simple.
2:36
Um, but I called two of my best friends
2:37
from Georgia Tech and we built this
2:39
camera. He said, "You have to have a
2:41
license plate." So, we did that, put it
2:43
up in my neighborhood, and about two
2:45
months later, we solved our first crime.
2:48
And that was the beginning of Flock. And
2:50
now, as you've heard, the company has
2:52
grown quite a bit since then. We do more
2:53
than cameras. We build cameras now. We
2:56
build drones. We build audio detection
2:58
devices. We build software that connects
2:59
it all together. And the concept is
3:01
actually still this simple 9 years
3:03
later. Car drives by. We do what we call
3:06
a vehicle signature. We look for the
3:08
make, the model, the color, anything a
3:10
detective would notice.
3:12
And then you catch the bad guy. And
3:16
that's what I thought we were doing,
3:17
right? And then I was speaking to a
3:19
police chief not too far from here in
3:21
Washington. And he said, "No, no, Mr.
3:22
Langley, this is more than about
3:24
fighting crime. This is as
3:26
transformational as DNA."
3:28
And candidly, I was confused. It's a
3:30
camera. What's the What's the rub? And
3:33
he said, "Well, look, let me explain to
3:35
you how poling used to work."
3:37
Historically, poling is a subjective
3:39
exercise. You drive around suspicious
3:41
neighborhoods looking for suspicious
3:42
activity, and ah, there is the root of
3:44
the bias. what's suspicious to me is
3:47
different than her and different than
3:48
him. That doesn't seem very good. And so
3:50
the chief was explaining to me that now
3:52
with Flock, it's objective.
3:55
That is a stolen car. That is a car that
3:59
has a child in the back of it. That is
4:01
the car that has an Amber Alert or a
4:03
silver alert. So we've moved the way
4:05
policing works.
4:08
It's pretty incredible. But let me give
4:10
you a real life example. So now we're
4:13
going to go all the way to Colorado. So,
4:15
it's a 911 call. It's a Levis outlet.
4:18
And and this might be the most important
4:19
thing I say all day. The 911 is
4:22
describing the getaway car as a white
4:24
van spray painted black and blue. Here's
4:27
a pro tip. If you're going to commit a
4:28
crime, do not spray paint your vehicle.
4:32
Do it in a Camry or Corolla. Something
4:34
simple, right? So, the 911 call comes
4:35
in. They have this van. So, they go into
4:37
this our product and search for a
4:39
vehicle. They can type in anything they
4:40
want. White van, spray paint, anything
4:43
that makes sense. Now, what's cool is it
4:45
turns every single camera into a
4:48
detective. Now, every camera in the city
4:49
is looking for this white van. The
4:51
second thing they do is they launch a
4:53
drone. So, just like Adam, launch the
4:55
drone up in the air. The drone's flying
4:57
at 400 ft, 65 mph trying to find this
5:00
white van. Now, a camera eventually
5:02
finds them. Drone over. Now, in a normal
5:05
reaction, as a law enforcement officer,
5:07
you have to assume the worst case
5:09
scenario. Is this person armed?
5:11
Probably. Is the public at risk?
5:13
Probably. So, you come in hot. But with
5:17
a drone overhead flying at 400 feet, the
5:19
suspect has no idea. Officers can
5:21
respond with precision and with
5:23
calmness. So, from start to finish in 21
5:26
minutes, from a 911 call to Levis's
5:28
outlet, tracking the car all the way to
5:30
a second outlet, there was an arrest.
5:32
Now, if you zoom this all the way out,
5:34
some of the biggest cities in America,
5:36
some of the biggest companies in America
5:37
rely on flock to keep them and their
5:39
employees and their citizens safe. Two
5:41
years ago, San Francisco was in a
5:43
crisis, as many of you probably know.
5:46
They deployed everything Flock has.
5:48
License reading cameras, drones, video
5:51
cameras, our software that ties it all
5:53
together. And Mayor Lur described this
5:55
fusion of sensors as a transformation in
5:59
public safety. And if any of you live in
6:02
the city, you know that it feels safe.
6:04
All of a sudden, crime is down 25%.
6:08
Homicide is the lowest rate it's been in
6:11
70 years. So, this all seems really
6:13
good, right? Well, while the company's
6:15
scale has grown and a lot of things have
6:18
changed for me, one thing's been very
6:20
consistent.
6:22
There is inevitably some part of this
6:24
audience today who's saying, "I don't
6:26
know. This snitching ass startup kind of
6:29
doesn't make sense to me."
6:32
And this was our first ever press
6:33
coverage. And I was I was so proud. I
6:35
framed this and my my poor mother was
6:37
like, "Why are you framing this?" And
6:38
I'm like, "Because it's a reminder that
6:41
when you work on important things, some
6:44
people will disagree and that's okay.
6:47
It means you should work harder." And so
6:49
when I talk to these opponents, and I've
6:51
I've got to know them many of them over
6:52
the years now, they tend to have three
6:54
big issues with flock. Whether this is
6:57
constitutional violation, what does this
6:59
mean to my privacy? And are you building
7:02
mass surveillance? These are pretty good
7:04
questions. Let's hit the easy one.
7:06
Constitutional. 30 court cases over the
7:08
last two years. Every single one of them
7:09
has unanimously said this does not
7:11
violate the constitution. Why? We're not
7:14
tracking a person's in real time. We are
7:16
tracking a fixed location. Not a lawyer,
7:19
but I'm going to go with the courts. So,
7:20
I'm going to check the box on. Second
7:23
one's privacy. This is actually a really
7:24
good question because nine years ago, I
7:26
did not have a degree in criminal
7:28
justice. I had no experience. I was an
7:30
engineer that just wanted to build a
7:31
product to help fight crime. But when we
7:33
sat there and said what if we are really
7:35
successful what could go wrong. So we
7:38
started looking what the industry did.
7:40
So the first thing we asked was what
7:41
about data ownership? Who should own
7:43
this data? This is a lot of data.
7:45
Historically companies my competitors
7:48
own the data and then they resell that
7:49
data for other use cases. In flock model
7:52
the customer always own the data. So if
7:54
you live in San Francisco it's San
7:55
Francisco's data not flocks. The second
7:57
was retention. Well how long should this
7:59
data exist? Historically, companies in
8:02
this industry stored the data in
8:04
perpetuity. That seems to me more like a
8:07
liability than an asset. So, we delete
8:09
the data at 30 days. If cities and
8:11
states have laws that say it should
8:13
store longer, we do that. Shorter, we do
8:15
that. The third is around
8:17
accountability.
8:19
With everything that's powerful should
8:21
come heightened responsibility. Every
8:23
action inside a flock is audited and
8:26
stored in perpetuity. to every action an
8:28
officer takes. And sadly, three officers
8:31
were arrested last year for abusing this
8:33
system. Now, I I think if anything that
8:36
speaks very poorly on the law
8:38
enforcement industry, but it is a
8:39
reminder that when you build tools to
8:41
hold people accountable, it starts to
8:44
work. And then the last one, and this is
8:46
the most interesting to me, if you go to
8:47
a place like the UK, there's 45 police
8:49
departments. You go to Australia,
8:51
there's eight. France, there's two. And
8:55
does anyone want to guess how many
8:56
police departments are in America?
8:58
You're way off. 18,000.
9:00
Duh. Okay, this is the problem. You
9:03
literally can commit a crime in city A
9:04
and drive to city B and they're like,
9:05
"Oh, who are you?" You leave state A to
9:07
state B, and they're like, "We have no
9:08
record of you." This entire problem is a
9:11
data problem. These cities have not
9:13
historically worked together. They moved
9:14
to the cloud less than 10 years ago, 5
9:16
years ago. And so one of the biggest
9:18
criticisms of Flock is that we create
9:21
this national network of data that
9:23
allows local law enforcement to work
9:25
together. It's a uniquely American
9:27
problem. Third question that I get,
9:29
right? So we covered the Constitution,
9:30
we covered privacy. Now the big one,
9:33
surveillance state. Is this a creation
9:35
of a new surveillance state? You know,
9:37
look, I think everyone in this room
9:39
probably has a cell phone. And what if I
9:41
told you that someone with mediocre
9:43
computer software skills could build a
9:45
product that tracks your exact location
9:47
in real time down to a few feet because
9:50
all of you use Google Maps and Uber and
9:53
any device that has access to location.
9:55
And so I think about the tooling that we
9:57
ask our local law enforcement to use and
10:00
we give them a weapon,
10:02
a gun, a taser, yet we're afraid to give
10:06
them a camera that reads a license plate
10:09
on a public road. I'm not saying that
10:11
it's not a choice. I'm just saying that
10:12
for me, the trade-off is I don't know
10:14
how many more 11-year-old girls need to
10:16
be kidnapped or human trafficking rings
10:19
or silver alerts or Amber Alerts we have
10:21
to go through before we say enough is
10:23
enough. Now the now the beauty of
10:25
America though is that every city,
10:27
state, county gets to pick. It's not my
10:30
choice. I live in Atlanta. I vote. You
10:32
should vote. The challenge is that when
10:36
you decrease surveillance and in effect
10:38
decrease safety,
10:41
it impacts certain parts of the
10:42
community more than others. If you are
10:44
affluent in South Africa, you have 9
10:47
foot walls. You have private security
10:48
guards. are 600,000 private security
10:51
guards in South Africa, more than the
10:53
military and police combined. So if you
10:55
are affluent in South Africa, it is a
10:57
beautiful place to live. You are crime
10:58
free. And if you are not affluent, crime
11:02
is simply the cost of being alive. Now
11:04
if you transpose that to a place like
11:06
Tokyo, beautiful city, there's a few
11:08
dozen homicides a year. Whether you go
11:11
to work at 2 a.m. or 8 a.m., you feel
11:14
safe. It just doesn't matter. And so
11:17
America is this like very interesting
11:20
conundrum in the middle. We have the law
11:23
enforcement infrastructure we need to be
11:25
safe. Yet we haven't made to me the most
11:28
important decision as a society which is
11:29
that safety should not be a privilege.
11:32
Safety should just be a right. It should
11:35
not matter the color of your skin, your
11:37
political affiliation, or your income
11:39
level. Everyone, in my opinion, deserves
11:43
the right to be safe. Thank you very
11:46
much.
11:50
>> It's so interesting persuasively argued.
11:53
But when you sell these systems to law
11:56
enforcement,
11:57
how
11:59
strongly do you feel you have the right
12:01
to limit what they can do? Because you
12:04
you've made a powerful case for these
12:05
things used the right way, but everyone
12:07
here can also imagine them used the
12:09
wrong way. Here's a protest. you know
12:12
that cameras do do things that someone's
12:15
cell phone doesn't do. You can detect
12:18
>> in in the end we're in a world where you
12:19
can detect individuals etc. Are you able
12:22
to restrict what they do or is it just
12:25
you sell the system and then it's over
12:26
to you act within the law my friends.
12:28
>> Yeah. So I actually want to answer your
12:30
question in two parts. The first is when
12:32
I I was having this debate with our team
12:34
because people were asking why do some
12:35
people dislike flock but not these other
12:39
things. I said you can see flock. I
12:41
guarantee you if you go to our website
12:42
and look at what our camera looks like
12:43
and you go back home if you live in
12:45
America, you will see our cameras
12:46
everywhere. They're easy to spot, but
12:49
your cell phone you just you're
12:50
oblivious to all the tracking that
12:52
exists online. So, you forget about it.
12:54
But to answer your question, um we do
12:56
have restrictions. Now, obviously
12:57
there's this thing called the
12:58
constitution which we care a lot about.
13:00
There's state law, there's local law,
13:01
but then there's things that we just
13:02
don't feel comfortable with. And so we
13:04
do set in kind of our own policy
13:06
framework and then just like most
13:08
technology companies, we have a full
13:09
team of attorneys who actually think
13:11
about this stuff full-time because our
13:14
customers are just not
13:15
>> that far the adoption curve.
13:16
>> I mean, how often do do you or someone
13:18
on your team hear of something that is
13:20
being done with your technology that
13:22
makes you uncomfortable?
13:23
>> Three times last year.
13:25
>> Like to me it it is horrible that in the
13:29
case was I'll pick on my own home state
13:30
because I live in Georgia. Uh, there was
13:32
a police chief in Northern Georgia that
13:33
used Flock to track an ex partner. It's
13:37
It's horrible. But then I'm reminded
13:40
that the GBI used Flock's audit log to
13:43
investigate that chief and he's now in
13:45
jail. And and so I think the thing I
13:47
think about to your question is like if
13:49
Flock did not exist, he was probably a
13:52
bad person. He was going to do this one
13:54
way or another. And so I look at the 1
13:56
million plus cases and go, okay, this is
13:58
a trade-off that I can make, but that's
14:00
my choice.
14:02
>> You've got such a powerful technology
14:03
and it's very cool to hear you open to
14:06
having these discussions. Garrett, thank
14:07
you so much.
14:08
>> Thank you very much. Thank you.
14:11
>> That was a heavy one, right? Like we're
14:12
all facing this question with ubiquitous
14:15
surveillance. And look, think about it.
14:16
It's totally legal to film in public
14:18
places, but now we have these AI enabled
14:19
cameras that are tracking every single
14:21
person or vehicle that drives by. And
14:23
then when something goes down, you have
14:25
a drone that from really high altitude
14:27
can zoom in and track a suspect. Like
14:29
this is insane technology that exists
14:31
today. This is obviously a controversial
14:33
topic and people are on different sides
14:35
of this. But at TED, we want to hear
14:37
arguments on both sides. Most people
14:39
want to live in a world that is free
14:40
from crime. At the same time, nobody
14:42
wants to live in a perpetual panopticon
14:44
where every move that you do is
14:45
surveiled. And a lot of the double-edged
14:47
nature of this technology comes down to
14:48
the people that are using the tools,
14:50
right? like who gets the keys to the
14:52
castle and how are they deploying this
14:54
technology. The best thing we can do is
14:56
draw awareness to these issues and have
14:58
a conversation around it because the
14:59
technology is one thing but the question
15:01
about how we utilize it in day-to-day
15:03
society is a collective conversation. So
15:05
regardless of where you fall in the
15:07
issue, we want to hear from you. So let
15:08
us know your thoughts. I hope you
15:10
enjoyed
