Original subtitles

[man] The term "pre-crime" comes from this movie,

Minority Report, in which a prediction is being made

about something an individual has not yet done

but is going to do, and a preemptive arrest is made

of someone before they've performed the act.

[man 2] If you would have asked me 37 years ago

if we would have gunshot detection or video cameras in neighborhoods

or be able to predict where crimes occur,

I would have said you're crazy.

[man 3] It's not aiming to a certain future,

the future is already in the present right now.

It's the securitization of our society.

[man 4] I have no idea what the next five or ten years

is going to mean to law enforcement in terms of technology advancement.

Just looking at where we've come so far,

it's gonna be mind-boggling.

[man 5] Can we predict an actual crime event before it occurs?

[waves crashing]

[man] When I heard of this story for the first time,

I thought, "Now it's finally happening."

Hollywood has eventually merged with real life.

Software that predicts where and when

the next crime occurs, police that arrive at the crime scene

before the perpetrator,

computers that generate lists with tomorrow's murders.

Pre-crime, they call it.

A friend writes to me, "There's something foreign

"looming on the horizon, and we can only guess

that it will shape us entirely one day."

[distant police sirens]

[man 2] Our strategic subject, list, or it's called the SSL,

is a system that we worked with a professor

from the Illinois Institute of Technology,

an academic partner here in Chicago,

to be able to assess and analyze those people

that are at the greatest risk of being a party to violence.

The system is able to prioritize and tell us

those individuals that we really have to focus on,

to work with to try to prevent that violence.

When you have so many different data sets,

or you have so many cameras to watch, which ones do we watch?

Social media, right? So many different social media communications out there.

How do you know what to concentrate on?

That's what our predictive aspect towards

that strategic subject list, that's what it does.

[man] They experimented with it a little bit in 2012.

2013 is really when it took off.

They came-- the police department came up with,

um, more than 400 names of people who fit that bill.

Individuals who were most likely to be prone to violence,

either as a victim or a perpetrator.

Each of the 22 police districts came up with 20 names.

And they were chosen, I don't know the science of it,

but it was all through mathematical algorithms, basically.

And it didn't have anything to do so much with

them being hardened criminals as much as it had

to do with, "Who are they arrested with?"

[mutters indistinctly]

[doorbell rings]

[dog barking]

-Mr. Robert McDaniel? -Yes.

Hi, I'm Commander West and I am with

the 15th District Chicago Police Department.

-May we come in? -Yes.

[Gorner] Like, take Robert McDaniel, for example.

He was not a hardened criminal.

He had been arrested for many minor offenses,

like gambling, shooting dice, or smoking marijuana.

Minor offenses, but the people who he was arrested with

during those crimes, some of them,

or least one of them, was a victim of violence.

So, the logic was is that, well, Robert belongs

on the list because he has a relationship with somebody

who's been a victim of violence,

because he's been arrested with that person before.

[man] I was unemployed, a school dropout.

It was either go get work or sell drugs in the street,

and selling drugs wasn't for me.

I tried to put myself...

I tried to get my GED. That didn't work.

And in the midst of me trying to get my GED,

I started getting followed home, started having police officers

walk up on me, ride up on me, saying my name, government name,

where I had been and... just things like that.

Then... I had Miss Officer West

and a social worker-- I can't remember his name,

but they had stopped at my home, they stopped at my house

and subpoenaed me, told me that I was put through some type of test,

they're saying I was supposed to be eligible to shoot somebody or get myself shot.

That I was put on the Heat List of 400 people in Chicago.

Mr. McDaniel, as part of our violent reduction strategy,

our software has generated a list of potential criminals, actors, and victims.

We are here today to inform you of the fact that

our computers have placed you on the Heat List of the police department.

Now, should you decide to continue to engage in criminal activity,

you should know we're gonna charge you

and we will prosecute you to the fullest extent of the law.

[McDaniel] I guess we was associated or had friends

that's supposed to be in the streets,

or however the scenario go.

But we was put through a test,

and we supposed to came out the most...

The, like, top 400 dangerous people in Chicago.

Now, yet again, like I asked you,

how can I be dangerous for smoking weed or shooting dice?

Who does this hurt?

[Caluris] The timeline shows all the criminal activity

that that person's associated with.

If you see on the bottom, those are all interactions

that he's had with the police.

Either as an arrest, as a contact,

as a victim, anything with it.

So, you know, who he hangs with, you know where he's been,

everything to do with him that we've documented

through police interaction.

Scroll down, please, whoever's got that.

This shows-- this is what they'll compile and put together

and get back out into the field within 15 minutes.

So, if this person is the victim of a shooting or a violent crime,

they'll pull up there, it's got their criminal history...

That's associates...

So everything you saw before was all the criminal history

involved with that individual, so there's probably

maybe about 25, 30 arrests that you saw on that subject.

The associates, people that they're documented

as having an affiliation with,

that's, again, a pretty comprehensive list.

Certainly, we can actually do even like a link analysis

to be able to show how that network interacts.

The idea that you could essentially connect

all of the data streams that government collects in different ways,

everything from your arrest records,

to your contacts, to your foreclosures,

to your mental health records, to your social benefits,

and put 'em in a particular computer database,

and then be able to do link analysis where you connect a phone number

from all the different sources and go out several links and be able to see the world,

is something you would never imagine that is technologically possible now.

[man] Let's summarize:

Firstly, they are quite serious about fighting crime with algorithms.

Secondly, Robert McDaniel is on the wrong side of the algorithm.

Thirdly, apart from its developers,

nobody knows how the algorithm behind the Heat List works.

Fourthly, in 2016, statistically,

2.0876 people are killed every day in Chicago.

How does an algorithm that promises safety actually work?

[distant police sirens]

[man] We've been using predictive policing

in Kent now for three years.

And we use a system called PredPol.

This is a system that is predominantly a patrol strategy for us.

So it tells our officers where the high-risk areas are today,

and they go to those areas and carry out their duties as they normally would.

We find the system is very adaptable,

very easy for our officers to use.

What we would describe as an operationally ready system.

We've helped to develop PredPol, which is an American system,

and we've played a large role in what PredPol looks like today.

In terms of the future, we're looking at how we can

protect against less tangible areas

around organized crime and modern slavery,

sexual crime, and protecting our borders.

And that's the next big challenge for us,

and we're working with PredPol around data fusion,

and also with the home office, the government in the UK,

to produce more and better predictive policing models.

[waves crashing]

[man] That's quite a remarkable list.

Organized crime,

human trafficking, sexual offenses, border security.

An algorithm that targets criminals just as migrants.

Question: what else does it target?

Question: who gets what kind of security?

Question: who is protected by the algorithm

and who isn't?

[Johnson] What's embedded within the algorithm, within PredPol,

is something called routine activity theory,

which is-- as you know, we to tend to do the same things

day in, day out, in the same way.

You know, I go to the same supermarket,

I tend to drive the same direction to that supermarket,

even though I don't have to. I tend to go to the same bar.

Um... criminals are the same, you know,

they're creatures of habit, and they don't like to take risks.

A lot of the theory and research within PredPol

is based on some pretty heavy criminological research,

most of it done in Europe that says, you know what?

There's a lot around repeat behavior,

repeat behavior of victims as well as criminals.

What we did initially with the system is give it,

it was put through five years worth of crime data

and three year's worth of antisocial behavior data.

So if you think about how long we've had it in place now,

that means we've got eight years of crime data in it

and about six years of antisocial behavior data.

We send data over on a daily basis

and the data is analyzed, and then that is sent back to us

in the form of Google Maps with red boxes on it.

Those boxes represent the high-risk areas

that we then say to our officers,

"Get in the box and police what you see."

[Gorner] They have like a ranking system...

which shows how many times more likely are they

than the general population to be prone to violence.

So, Robert had a rating of 215,

which meant he's 215 times

more likely to be prone to violence.

But Robert wasn't... I mean, that paled in comparison

to a number of other people on the list.

There were a lot of people on that list

who were than 500 times more likely

to be party to violence.

And, again, that's not because of their criminal history.

That's because of the people they've been arrested with.

[man on P.A.] Doors closing.

[Ferguson] You know, so, this is what's really frightening,

is that there are companies now scoring every citizen.

And our political elections right now,

the Republicans and Democrats are really doing

this nanotargeting where they really are scoring

families to the address, right? So that information's out there.

It's not really out there whether they're a felon or not,

but it is out there.

And so what the police here are doing,

is they're literally just purchasing the information

that other people already have.

Now that scored society, of course, is frightening.

Of course it's not just a privacy sense

that you are giving out this information.

It's about a government owning this information, right?

This is a different situation.

In America, we are pretty willing to let big companies

like Google and Apple know pretty much everything about ourselves,

but we're more reluctant to have government, and that's a good line to draw.

What's happening here is this sort of data convergence,

where you're really seeing private companies

collecting this information, and then essentially selling it

or offering it through different services to law enforcement.

And people do know that information.

Even if the police don't have it,

the private companies do have it,

and that's part of where we are now

as technology is collecting as much information

about us as it can.

[man] Facebook like, click, saved.

We deliver the data which generates information about us

that circulate on the internet until the next update

creates yet another data set, and so on.

Data mining endlessly.

Somewhere fairly lost at the bottom of the digital food chain,

there are people like Robert McDaniel.

Privacy? What privacy?

[man] Whenever someone fills out an application for a loan,

they're going to be providing certain information.

The Beware Software, they access databases

from financial institutions, the courts,

any type of loaning institutions.

Beware has the ability to access all of those databases, simultaneously.

So when a call comes in to our dispatch center

and it is categorized as a life-threatening call or a in-progress crime,

then-- and there's an address attached-- the Beware Software

automatically searches all of these databases

and then provides the operator in the real-time crime center,

information specific to that address.

People that have lived there, do live there,

their cell phone numbers, prior addresses, associates.

The other piece that Beware allows for

is to research social media

and to gather any type of information

that might be in there in terms of threats.

The theory behind Beware makes a lot of sense.

If I was that police officer on the street

and I was entering a house and I didn't know who lived there,

I'd want all the information I could.

But the problem is, if it's sourced through these data brokers,

there just isn't really much accuracy.

So, you might be arriving at a house, and the address alerts

says "a dangerous place," that maybe somebody lives there.

Maybe that dangerous person moved.

The problem is accuracy. I mean, have you ever gotten

the wrong catalog in the mail and, like, "Why did someone send me a particular catalog?

I don't have children, why do I have it?"

That's the inaccuracy that comes along with these

data brokers, right?

They don't need to be perfect,

because what they're really doing is trying to sell products to people.

Well, if the algorithms used in the private sector

have allowed them to become more successful

in targeting their audience to sell product,

then we should take advantage of that same algorithm

that allows us to become more successful

in law enforcement and preventing crime.

[people chattering]

[man] In the case of the Beware Software,

I think the bad far outweighs any potential good.

And I can see how in a perfect world,

and if the software were perfect,

it could help make police officers safer.

The problem is, nothing is perfect.

For instance, one of the things that the software

company says that it looks at

are postings on social media, such as Facebook and Twitter.

There was one woman in another city who was flagged in the software

for making comments on Twitter about rage.

Rage has a very specific meaning in terms of anger,

violence, but the rage she was talking about

was a card game called "Rage" that had nothing to do

with violence, or aggression, or anything like that.

Yet she was flagged as being a possible problem

because she had sent these messages about rage.

And what if someone is making comments about that they don't trust the police?

Is that gonna flag them as being a potential problem?

So there are too many opportunities for the computer to get it wrong.

And if they get it wrong coupled with a police department

that already is much more likely than other police departments

to shoot citizens,

that's a recipe, potentially, for disaster.

[waves crashing]

[man] Taking up one question again.

Why are we forcing these technologies upon ourselves?

The Silicon Valleys of this world are making

a lot of money with them, okay.

We, the users, enjoy the comfort in Google land, fine.

And that's it?

What if the internet, fed by the permanent feedback of its users,

already had its virtual awakening?

What if it developed its own needs and interests?

If it was always leading us to more convenient technologies

because we pay for it with private data?

What if freedom was just an illusion?

[man] If we build a game, a city game,

in which technology's in everything and you're a hacker,

then you'll have a lot of control and a lot of power.

So that was an interesting idea for us.

So I started doing research.

And we started looking at what existed already.

And we started seeing that a lot of that technology was already out there.

That it was technically almost feasible to do that.

So instead of us trying to invent science-fiction,

we started grounding our creation to reality.

You know, we've all heard of big data,

the notion that all those cameras,

but also all those devices-- our phones,

computers we use, the chips in our cars--

they're gathering our informations.

Those informations are stored. We don't even know where.

Oftentimes we accept that they get accumulated.

And big data, it's a neutral technology.

It can be used for good, it can be used for bad.

We like calling it the digital shadow of people.

You knew what was their earnings...

you know, some important facts of their lives.

And as you go around the game simulation,

you could open what we called a profiler,

which would tap in all those devices,

do facial recognition and all that,

and give you information on the citizens

of our virtual Chicago.

This is not a dystopian game.

It happens in a little different version of our world.

It's very much grounded in reality,

but we turn the dial to 11

and we say, "What if this happened?"

And so it's not the dystopian reality,

it's before the dystopian reality,

when things can be done about it.

[man] "Metropolitan Police, territorial police,

"are working together for a safer London.

"To who it may concern, the Metropolitan Police service

"and all of its partners are committed

"to reducing knife and gang crime.

"Information indicates that you have, or are associated with,

"a gang that is linked to crime.

If you are involved in crime and do not stop..."

"If you are involved in crime and do not stop,

"you may be targeted by the police and partner agencies.

"Under a piece of legalization called Joint Enterprise,

"you may be convicted of a crime and sent to prison

"for just being present when a serious crime is committed

"or being with those persons who commit a crime

"and you don't try to stop it.

"You will need to change your lifestyle.

We can help you to do this."

[chuckles]

"You can speak in confidence to a police officer

"and/or contact any of the organizations listed at the end of this letter.

"I would encourage you to speak to them,

"as they can help you break any gang links.

Yours sincerely, borough commander."

That's some bullshit.

Reading that letter just took me back

to when I first received it.

And it's funny, 'cause everybody that's in my circle,

yeah, we all received that at the same time.

You can't ever, ever let somebody tell you

that you can't chill around somebody,

you can't go to a certain area, you can't do certain things.

Especially when these people are trying to bring us down.

They put things like this into place just so that

we can act up and-- do you know what I'm saying?

Because what they don't realize, yeah--

actually, no, they do realize it.

The police do realize it.

They realize that we haven't got a voice.

Do you know what I'm saying?

It's like when you was in school once upon a time

and you put your hand up to ask the teacher something

or to answer a question, and the teacher don't point at you.

And you get frustrated, do you know what I'm saying? You get frustrated.

The teacher don't notice you until you get mad.

That's the only time that they notice you.

So that's how the mentality of us...

The only way that we can get our point across

is by putting fear into people because that's the only time

that they're willing to listen to us.

[man] The doctrine of Joint Enterprise

was actually brought in over 200 years ago

to stop people encouraging duels.

So, if two people were dueling, whether it's by pistols or swords,

their seconds, who are there to support them,

they can be done for joint enterprise

if someone's killed.

So that doctrine is not actually law.

It is doctrine adopted by the courts.

But it has an operational and tactical implication

in terms of the Matrix.

Um, the Matrix wasn't around when I was in the Met.

I retired three years ago.

But the Matrix is primarily an enforcement tool

to be more proactive,

to target those who are committing the crimes.

Unfortunately, when you cast the net very wide,

you are bringing in people who are not necessarily involved in gangs.

[man] For whatever reason they called their database "Matrix."

The name, however, seems like a bow to a technological system.

The London Matrix works like the Heat List in Chicago.

Identify individuals, connect them,

detect patterns and social networks,

calculate the statistical possibilities.

Score people, issue warnings, keep an eye out.

Question: is there an algorithm targeting corporate crime?

[man] They sit there, they just wait, they just watch.

They have unmarked cars on the block.

And they post up on the sides somewhere they can't be seen,

or they will have officers on foot in plain clothes

to watch what we're doing, like walk around the area,

see what's going on and that.

They know who you chill around, they know who you're around,

they know what you do 'cause they're watching,

do you know what I'm saying?

You can't really escape it.

I'm doing an anti-gang project in a local area here in East London.

I have clients who are saying, "I have never been involved in a gang."

But the real issue is the subjectivity to get people

on the criminal intelligence system, the criminal system.

And then how that then goes into the Matrix,

to then associate people in certain gangs, which are questionable.

So that the thing is who's checking the data entry?

Who's checking those officers who commit those data entries?

[man] I don't see myself as a gang member.

I'm not a gang member.

I'm not a gang member, do you know what I'm saying?

I'm just a part of a circle that...

I'm just a part of a tight circle which always

gets shined a dark light on, do you know what I mean?

And always will by them.

[man] If you tell them, "I'm not in a gang,"

they'll ask you, "What community are you from?"

"Oh, I'm from the [inaudible]."

"Okay, those are Body Snatchers over there.

Those are Four Corner Hustlers. You a Four."

How can you tell me what I am because of my address, because of where I stay?

That's probably where I can only afford to live. That makes no sense.

I just honestly believe they got a job to do, and they want to do it.

If they gotta breed crime, if they gotta make criminals,

if they gotta sit here and convince you that you're a criminal

and provoke you to do it, they'll do it.

Actually, I had a friend killed a couple months-- couple weeks prior to that.

Only part I can say is, they labeled it a gang murder.

Now, I guess that's how I got affiliated,

'cause me and a person that was murdered was so close.

But other than that, I actually don't know.

But like I said, I ain't did nothing that the next kid did.

[woman] They haven't told us what the algorithm is

that they're using to identify people.

They haven't told us what that data is.

And there's no way to get off the list,

that we're aware of, once you're on it.

So, that's scary to a lot of people.

It's frightening to not know how the list is created,

or to be able to get off of it on the back end.

And they can say that, you know, we're using math,

we're using science as a way to do it.

Math and science aren't always right.

There's been reporting on these algorithms

that are used in sentencing after trials,

where people are given a number to predict whether or not

they're gonna be more likely to engage in crime in the future.

And that impacts their sentences.

There's been recent reporting showing that those algorithms

aren't accurate, and that they have a disparate impact on the basis of race.

So, that's a danger of these algorithms

and one of the reasons that we want the police department

to be open about the factors that they're using,

and the algorithms that they're using in this case,

because it should be put to the test.

[man] So, a police department first uploads their data.

This is actually open data from Philadelphia.

It's about 800,000 records.

And a very important point in this software

is telling the system how important each type of crime is to prevent.

So, the harm created from a homicide is equivalent

to $8.6 million in comparison to, let's say, the robbery,

where it's like $67,000.

So, the system needs to know that so it can know

how to balance the use of resources, right?

Is it more important to prevent a very frequent low-level crime,

or a very infrequent high-level crime?

And how to balance those things.

And this is what basically defines that.

So, that's all in the administrative interface,

and then the officer just logs in

and then sees this map of, in this case,

the entire city, and officers in Philadelphia are assigned

to particular patrol service areas. which are the black boundaries here.

So, they look at the map and zoom in on the beat that they're in,

and they see the boxes, and then the color of the box represents what is the focus.

And the tactics is the recommendation from the algorithm, the model?

From the software, yes.

Which means that in some years,

the machine learning tools will learn which kind of tactics

work with which kind of situation.

[man] Exactly, yes, yes.

All data is biased, but police department data,

incident data, has the potential to be biased in a number of different ways.

And we cannot eliminate that bias,

but we can potentially offset it to some extent

by incorporating other data.

We have a number of different components to the software.

One is a planning component

that someone can use at the police station.

The second component of the software is a mobile version of the software

that can be in a car or other vehicle, and the police officer

is able to see as the car moves around,

are they inside one of these priority mission patrol areas.

It's actually using the GPS from the tablet

to track our location, and as we enter boxes

it's going to update the display with information about them.

So we actually are just driving through a box right now, which is about robberies.

And if this was our final destination, we would start patrolling

for about ten to 15 minutes in this area.

It's still relatively unlikely for any crime

to happen in that location at that time.

It's just that this is the highest-risk location

amongst all the choices that we have available,

and so it's the best place for the officer to spend that free time.

While we are positioning an officer in a particular place,

which means that they're going to be paying attention

to that place, that should not give them the authority

to assume that anyone in that place is a criminal

unless they see something that's actually criminal in nature.

We have generally been very cautious about

incorporating any kind of person-centric data into our models.

We believe there's a number of substantial problems with this,

whether it's privacy concern

or just the accuracy of the actual modeling.

We're not using surveillance data in HunchLab,

but I think it's a key question that our societies need to be asking,

and under what circumstances is it reasonable

to take advantage of that kind of data?

[man] Big data, we the users, and our privacy.

Well, who could have imagined years ago

that Google can algorithmically calculate what I will do tomorrow?

Simultaneously, we activate things that were silent until now.

Everything that once was quiet starts communicating

with the world and sending our data to the internet.

My toothbrush, my TV set, the chip under my skin,

my fitness tracker, the toys of our children.

[man] I was not aware of these kind of "technologies,"

quote unquote, being implemented by the police, et cetera.

It's not particularly surprising, because the technological developments

in terms of policing, be it domestic or globally,

is developing all the time.

And it's something that we're all privy to,

we can all see it on our TV screens,

especially when it comes to foreign policy

and conflicts that the West are conducting abroad.

How do I feel about it?

I'm really concerned, because I work with

a lot of young people and young adults and children

who are, or have been, or will be, unfortunately,

in the short term, most likely be involved in the criminal justice system,

because they come from troubled backgrounds,

or they're working class and they're black people.

So, if you can use some kind of predictive technology

and software to predict,

according to what these people do in the future,

it's not gonna predict anything particularly positive for them.

But if you want to make money, software to have algorithms

to give to the police, it's...

it's indicative of how our society

is progressing away from human solidarity

and a human approach, to just squeezing people

as hard as you can in any and every which way.

[man 1] That's one camera over there.

[man 2] Yeah.

[man 1] Then there's one camera over there.

There's one camera over there.

-You see her? -Yeah.

And then there's another camera just back there, you see?

-Yeah. -Over there.

And so three murders in the last year in the park.

So, what the fuck are these cameras doing?

Yeah, if they're not being used to solve murders,

then obviously they've got some kind of different purpose.

Why are they here?

They hit a tick box, they say to people, "Oh, we're doing something."

Yeah. And if you look at these cameras,

they're not the kind of ordinary CCTV cameras.

-They're sophisticated. -Definitely.

[man 2] There's some kind of box attached to it.

When you think about, three murders happening here

right as kids play... it's crazy, man.

This is where my son has played the last 11 years of his life.

And that's why more incidents happen here than in do in, like, gated parks.

that you can't get in.

If you think about how small this park is and how many cameras there are,

they've got full coverage of it.

And so, you know, where these gang murders are happening,

it's weird that they're not able to prevent them

with all the intelligence they have.

Stingray, triangulation, all this kind of stuff,

I'm not surprised that it's on this scale,

and I think there's actually, we probably don't know

most of the kind of surveillance abilities they have.

I think it's interesting that maybe some of the way

that they experiment on gangs and the black communities

is also sort of being used in political protests

and political organization as a way of sort of pioneering it and developing it.

I mean, of course, it's just reflective of the way

in which police mainly target black males

when it comes to crime and how they're disproportionately

stopped and searched, how they're hard-stopped.

[man] I could tell you about a time where I was

in my auntie's square, just walking.

I was literally just going to see my auntie.

I was walking, and then I heard them.

You know the slider doors on the van?

I heard the slider door open.

I wasn't really paying attention to what was behind me,

but I heard that open and I heard footsteps.

As soon as I turned around, boom, I'm being tackled to the floor

and getting my face kicked in by four different officers,

one female, three males, do you know what I'm saying?

Obviously, they rushed me, they jumped me.

I couldn't... I'm so excited,

So I couldn't even see faces. I just started swinging.

I punched a couple of them.

At the station, they said if I admit to hitting a police officer

they're gonna drop every other charge.

And the other charges was,

I matched the description of a robber.

Young... [indistinct]

hooded, blah, blah.

This is Tottenham, breh. This is Tottenham.

What else are you gonna see

besides young [indistinct] in hoods?

Hoods make us feel comfortable. This is how we dress.

This is just the culture, do you know what I'm saying?

There's a million other people... why me?

And I resisted arrest. I resisted arrest,

so they say they're gonna drop all the other charges

if I admit to hitting a police officer.

This is what they do. They're all about setups, man.

They were on the block watching me, do you know what I'm saying?

They saw me walking down, they did what they did.

And you know what? Police are the biggest gang in the world,

so they got a cheek.

[Logan] We, in London, and the other parts of this country, are a police service.

So, you've got to know what that means.

That means accountability and transparency

in all of your processes and practices.

Now I challenged that when I was in the Met,

when I was chair of the Black Police Association.

I also gave evidence to various inquiries

that said the police service was institutionally racist

because of the way in which they conduct themselves.

Now the Matrix, for me, is another form of institutional racism.

It is racial profiling. It is unaccountable.

And, as far as I'm concerned, there has to be a way in which you can get off this system,

whether it's the Matrix, the DNA database, you know.

You've got to have a process where people believe

that the police service can be held to account.

[man] Matrix, strategic subject list, no-fly list,

selectee list, terrorist watch list.

Once on the list, always on the list,

because the computer says so.

Because, due to the algorithm, nobody is directly responsible.

Because there is no regulated procedure against the errors of the machine.

Because, let's be honest, nobody cares about

what consequences the decisions of a program have

for the life of Robert McDaniel, or Smurf.

[Guay] Our main character actually gets unjustly profiled

by the crime prediction system in the game,

and very early on the player wants to wipe that prediction

on his person, because it's unjust.

He doesn't feel it's justified.

It thinks it's arbitrary

that some code decided that he's a criminal

because of this and that reason.

So we're kind of twisting the perspective

from the being the vigilante,

now you're potentially the unjustly profiled

by the algorithm.

[sirens]

Codes don't have a conscience, and people who write code,

they don't always know exactly what's gonna happen afterwards.

There's a part of it that is beyond the moment where it's being engineered.

And so, since code doesn't have a conscience,

and the programmer doesn't always realize all the repercussions

of what he's programming, and he doesn't know

the extent of the data it's gonna be treating in the decades afterwards,

there is a point where, who's responsible, really, for the code's decision?

[Ferguson] If the Roberts of the world are walking down the street

and police know who's on the Heat List,

they might get stopped easier, might get searched more often,

they might really feel like they were living in a true police state.

And from the police perspective, they're like,

"No, we're just targeting the worst of the worst.

This person happens to be on it."

The negative way of looking at that

is probably closer to the perspective of the young man

whose door is knocked on by a detective who is there

to essentially inform him that there is a new surveillance

with his name, and in fact they'll give him a letter,

a custom notification letter with his name on it, with his prior record,

and with the future consequences should he mess up.

A very directed, very personal message of,

"We are watching you, we are surveilling you,

"we are giving you a chance to avoid violence

"or to be caught up and be arrested.

And you need to choose wisely."

[McDaniel] If I was to commit a crime or a murder right now,

they'd parade themselves.

Medals, honors. "Oh, we was right, look, our test."

Like, that's what I think they're looking for.

They got a grant, they made up a theory,

they started a program, and now they want results.

For whoever gave their money to fund this,

they have to give results, 'cause if you don't give results, he'll end it.

It's wrong. It's wrong to be profiled.

It is wrong to be saying that you something that you're not.

It's wrong for somebody to tell you you're a killer.

Like... I don't know.

[Gorner] The police say, "Hey, it's just another tool on our belts,

"it's something innovative, it's something that we figure

can revolutionize law enforcement."

But the flip side to that

is just from when I was doing the story in 2013,

is that a lot of the names on that list are young African-Americans.

So in the African-American community,

there were some rumblings that,

"Oh, well, this list is just another example of racial profiling."

[Hill] We don't have the personal capacity

to keep in touch with millions of people,

thousands of people.

Social media does.

And it's a double-edged sword

of, how do we try to navigate that?

And I think the state and the police are fully aware of this.

I think, ultimately, they're gonna be one step ahead.

They're gonna know exactly how to batten down,

or how to limit, for example,

thousands of people attending a demonstration.

"Oh, let's just cut off all their invites.

"Let's just not let it be imprinted on social media.

"Let's affect the algorithm to who is sending what messages,

who is posting what," right? There's no...

For us, we use it with the intention of,

"We don't care." They're gonna know anyway,

so we organize openly.

But, practically, I think it's a really worrying thing,

because we have no way to monitor what they're doing,

and we have no idea of the scope in which

the powers they have, and the technology they have

in terms of mass surveillance and databases.

[man] What we started with was a man found in an alleyway shot, no witnesses.

So that's all the information we had to start with.

What we were able to do is find the crime,

go back in time, see what happened.

These three cars right here, are all involved in this murder.

We do what's known as "wide area surveillance."

So we watch major cities at a time.

We take a very high-resolution image of a whole city

every second, process it to make it look like Google Earth,

and it allows us to go back in time and see

what happened within a city up to about real time.

Well, because we capture a major part of the city

at enough resolution, we can look at a location

where a crime occurred,

that we didn't know it was going to occur,

see what happened and follow the vehicles and the people

to and from the crime scene.

It is a very powerful tool.

This may sound like science-fiction,

but it's pretty straightforward.

So, what we're gonna do is, I'll just show you here,

this is a city, we're flying at about 13... 12,000 feet.

And we're able to zoom in anywhere within that area.

And what we'll do is we'll see when a female police officer leaves her house.

She is right here.

This car here is a lookout car watching her.

And then there's three other vehicles right up here.

So, here she actually leaves her house right here.

And immediately following, the blue car takes off

and tries to catch up to her.

She slows down and then rounds the corner, and then guns it.

She gets around this corner before the blue car gets around that corner.

And then she gets around this corner, and the blue car actually gets lost.

Unfortunately, these other three vehicles up here saw her,

and they're actually able to catch up to her.

So, here she is, she's slowed down by another vehicle.

And she gets caught behind this.

You get a vehicle that just backs out in front of her here.

We couldn't tell if he was directly involved or not.

But, unfortunately, that gave these other guys time to catch up to her.

And they catch up to her right here.

And she is shot six times in the head and shoulders.

And she's actually gonna run right into that parked car there.

And that is where she is found by the police in a few minutes.

And then what we do is we follow those vehicles as they flee the scene.

And here you can see they're passing people on the right and on the left.

Now we're actually able to follow two of those vehicles to final end points.

So within a few minutes after the murder,

we actually have two locations identified.

Then what we do, is we come in

and we'll actually go into Google Earth

to actually identify the houses,

and feed that information into law enforcement.

And we'll actually get the front door picture

of the house that he went into.

We'll take that image and send it forward to the police officers,

so they know which door to knock on.

[man] At some point I have stopped thinking about

who might know and who might store what about me,

why, and since when.

Probably because it doesn't make a difference anyway.

It's like with Hollywood, computer games, and television.

Everything is inextricably interwoven with reality.

Reality being just the medium of the basic code.

Zero, one, like, don't like, buy, don't buy,

guilty, not guilty.

Just to remind you, code has no conscience.

You know what was fun?

Like I told you, I lost a best friend.

When I say a best friend, a brother.

I watched them sweep his murder under the rug

like it didn't mean shit.

But it got put through a $2 million...

test that said I was a killer.

[cell phone rings]

[groans]

My friend was 18 when he got killed.

He got a son right now that I still look after.

That little boy actually think I'm his dad.

And I don't know.

It's wrong. It's just...

it's too much.

[man] If policing is going to use software to predict

what these people do in the future,

it's assuming that certain people with a certain history

are going to do certain things.

And that's just not necessarily the case,

because humans can change according to what support

and what personal decisions they make according to that support.

But, again, it's a metaphor for that, but it's also... it's what policing is.

Policing is not preventative.

It's not just in any kind of way.

And so it's just there as a punishing mechanism.

As a criminalizing mechanism, and as a punishing mechanism.

[Smurf] You know what's funny, yeah?

There's a funny thing, man...

A little story that's quite disgusting, actually.

Because you see how they're restricting so many things

that we're able to do and not able to do.

Music is one of them. Music is one of them.

'Cause I've got a cousin who knows people

who did a crime where someone got murdered,

and he made a music video.

And in his lyrics,

he was mentioning stuff about the crime.

Even though he didn't commit the crime,

he mentioned something about it. And do you know what?

He got sent to prison for six years just for the lyrics in the video.

And he got banned from rapping. Do you know what I'm saying?

Because, obviously, music is our way of expressing ourselves.

That's the only way that we know how to express ourselves and communicate, really.

Do you know what I'm saying? Just to express yourself.

And that's... that's... I think that's disgusting.

They're leaving us with no other option.

I've been told they started off a couple months ago with 400, but now skip to 1,500.

That's a big number.

You just made 1,100 more criminals right there.

So, now, who is that good for?

Is it good for the streets or is it good for the police?

You got more criminals to lock up. You got more cases to solve.

You got-- see what I'm saying? You got more product.

[sirens] [radio chatter]

You want to build a system, build a system that helps me get up out of this.

Help me go to school, help me go off and start a business.

That's what I say, like, even with this pre-crime stuff,

I try to just stay focused, work, stay out of harm's way,

stay away from anything that's negative.

And just try to prosper, try to be better than I was yesterday.

I could have let it turn me into thug.

I could have said things, just went off

and let people's images change me.

But to me, personally, that's not me.

I'm not gonna let you change me because you got

a theory about me, or that's just your personal view.

At the end of the day, I can always prove you wrong.

[Caluris] We're on our third version right now.

The system, we've been in place now for about three years,

so we've done regression studies to look at

the different variables we put in

and be able to test it out over past data sets

to make sure that those factors that we're weighing are effective.

However, I don't think we're ever gonna reach

a final conclusion with the program or the initiative,

'cause we're continually looking to re-evaluate it

and be able to add in additional data

and additional variables that could play

an additional significant role towards positive outcomes.

[distant police sirens]

[Gorner] That's just an example of the efforts that our city

is trying to make to fight the violence,

just trying to come up with creative ideas

in fighting it, and are there potential civil rights violations there?

Well, there's always that concern.

Um, especially, like I said with this history,

with the deeply rooted distrust between the police and minority communities.

See, my story might not mean shit to nobody, because it wasn't you.

But what about when they got your son,

when they got your daughter, your child... when it's you?

Now it's a problem.

Now everybody want to make it seem like okay, let's kill.

It don't affect nobody until it come knocking on your door.

[Ferguson] So right now I think people are willingly giving up this information.

Not just what you're giving up on the internet,

but as we move into a world of the internet of things.

Your smart house will reveal when you've left for the day,

when you take your shower, what temperature your bath is.

And your television can listen to you,

your car will be able to monitor where you're going

if you have like an OnStar system that tells you where to go.

Your cell phone knows all those things and what you're doing

and the conversations going on.

Like, we're just giving up this data to private companies

in a way where we're not really thinking about the consequences.

We're not thinking about what these data trails mean.

And for law enforcement, you could see just how valuable that information would be.

Why do you have to drink cold coffee in a hot car

surveilling some guy, when you could just use the internet of things

to track them all the way through, right?

This is the new world. And right now, the policy makers

and even the lawyers, haven't really thought through the consequences.

They haven't figured out how does the Fourth Amendment adapt?

How do privacy laws adapt?

How do laws that we have about telephone technology

apply in a world where suddenly your watch

is talking to the world and giving them your heartbeat,

and the rest of it, right?

We just haven't figured that out yet.

And it's important, I think, to ask these questions now.

I think we're at the very beginning of a very big conversation

about what we should do with this new data.

[man] Robert still has a score of 215.

Smurf has left Tottenham.

The internet is ever-learning and evolving.

Observes, surveys, collects, saves.

"Above the city," as a friend wrote to me,

"the sky is the color of a television

tuned to a dead channel."

Time to say farewell and go back home.

Back to my smartphone, my IP address, my emails,

my bookmarks, my Twitter account, my Facebook timeline.

Welcome to the Matrix.

[rock music]