AI Crime Mapping: 7 Real-World Examples

AI crime mapping uses smart computer programs to help police work better and keep communities safer. Here's a quick look at 7 real-world examples:

1. PredPol in Los Angeles: Guesses where crimes might happen
2. ShotSpotter in Chicago: Finds where gunshots come from
3. Hikvision AI Cameras in South Africa: Watches for crimes in real-time
4. Domain Awareness System in New York City: Gives police real-time info
5. Cloud Walk in China: Tries to guess who might commit crimes
6. HART in Durham, UK: Guessed if someone might commit crime again (no longer used)
7. Cape Town's Gunshot Detection and Drone System: Finds gunshots and watches from above

These AI tools help police:

• Find crime patterns
• Guess where crimes might happen
• Use their time and resources better

But they also cause worries about:

• Privacy
• Fairness
• Data safety

Example	Main Feature	Does It Work?	Main Problem
PredPol	Crime prediction	Mixed results	Might focus on certain areas too much
ShotSpotter	Gunshot detection	Often wrong	Expensive, might not be fair
Hikvision Cameras	Real-time monitoring	Helps police respond faster	Privacy concerns
Domain Awareness System	Info sharing	Improves police communication	Hard to maintain
Cloud Walk	Behavior prediction	Unclear	Big privacy worries
HART	Reoffending prediction	Stopped using it	Wasn't fair, hard to use right
Cape Town System	Gunshot and drone monitoring	Helped catch some criminals	Needs quick police response

As AI crime mapping grows, we'll likely see smarter computers, new tech working together, and efforts to make these tools fairer and more transparent.

Related video from YouTube

Main Parts of AI Crime Mapping

AI crime mapping uses smart computer programs to help police work better. Here are the main parts:

Data Collection and Analysis

AI crime mapping looks at lots of information from different places:

• Old crime records
• Social media posts
• Weather and area info
• 911 calls
• Traffic reports

The computer quickly finds patterns and spots where crimes happen most. This helps police plan better and stop more crimes.

Predictive Modeling

AI can guess where crimes might happen next. It does this by looking at old data and what's happening now. Here's what it can do:

What AI Predicts	How It Helps
Crime Hotspots	Shows where police should patrol more
Repeat Offenders	Helps catch people who keep doing crimes
How Crimes Happen	Helps police stop crimes before they start

This helps police use their time and people better to keep everyone safe.

Real-Time Monitoring and Response

AI crime mapping also helps police right away:

1. Sorting Emergency Calls

The computer decides which 911 calls need help first. This means police can get to the most important calls faster.

2. Watching Traffic

AI looks at traffic to find crashes and jams. This helps police and ambulances get to emergencies quicker.

3. Finding Faces

AI can spot people's faces in videos. This helps find suspects or missing people faster.

Checking Police Work and Managing Evidence

AI also helps make sure police are doing a good job:

• It looks at arrest numbers and how happy people are with the police.
• It helps sort through evidence like pictures and DNA faster.

This makes police work better and solves cases quicker.

1. PredPol in Los Angeles

Technology Used

PredPol is a computer program that helps police guess where crimes might happen. It was made by a UCLA teacher and the Los Angeles Police Department (LAPD).

How It Works

The LAPD has used PredPol since 2011 to help stop property crimes. Here's how:

• The program makes daily maps showing 10 small areas where crimes might happen
• Each area is about 500 feet by 500 feet
• Police check these areas when they can
• GPS in police cars tracks where they go

The LAPD now puts these maps online and on social media. This helps people see what the police are doing in their neighborhoods.

Does It Work?

People aren't sure if PredPol really helps:

Good Things	Not So Good Things
LAPD boss thinks it helps	Some police departments stopped using it
Better at finding crime spots than people	Might make police focus too much on some areas
Helps police know where to go	Only works if police use it right

Problems

1. Old data might make police focus on the wrong areas
2. Police need to use the program the right way for it to help
3. People might worry about their privacy when police share crime maps
4. It costs money, and some think it's not worth it (LAPD paid $50,000 for it recently)

2. ShotSpotter in Chicago

What is ShotSpotter?

ShotSpotter is a system that listens for gunshots. It uses sound sensors to find where shots come from and tells police quickly.

How Chicago Uses It

Chicago started using ShotSpotter in 2007 and added more in 2018. Now it covers:

• 12 police areas
• About 100 square miles of the city

The system works with special police centers that watch for crimes happening now.

Does It Work Well?

Good Things	Not So Good Things
Finds shots police don't know about	Often thinks other sounds are gunshots
Police get there faster	Costs a lot of money
Helps find evidence	Might make police go to some areas too much

People aren't sure if ShotSpotter really helps:

• Police say it helps them find people who've been shot
• But a study found that most alerts (89%) didn't find any gun crime

Problems

1. Not Always Right: It often thinks other sounds are gunshots
2. Focuses on Some Areas: It's mostly in Black and Latino neighborhoods, which might not be fair
3. Expensive: Chicago paid $33 million for 3 years, which is a lot
4. Privacy Worries: Some people don't like that it's always listening

Even with these problems, some people think ShotSpotter helps police stop gun violence and get to shootings faster.

3. Hikvision AI Cameras in South Africa

What Are They?

Hikvision, a Chinese company, put smart cameras in South Africa. These cameras can:

• Recognize faces
• Spot odd behavior
• Watch for crimes happening now

Where Are They Used?

The cameras are in big cities like Johannesburg and Cape Town. They work with:

• Old security cameras
• Control centers where people watch the videos

Do They Help?

The cameras seem to help in some ways:

Good Things	Not So Good Things
Police get to crimes faster	We don't know exactly how much crime they stop
It's easier to find suspects	Some people worry about privacy
Criminals might think twice before doing bad things	The cameras might treat some groups unfairly

Problems

Using these cameras isn't easy. Here are some issues:

1. Privacy: People don't like being watched all the time
2. Unfair treatment: The cameras might pick on some groups more than others
3. Foreign tech: Some worry about using Chinese cameras for safety
4. Money: The cameras cost a lot to buy and keep working

South Africa is still trying to figure out if these cameras are worth it. They're looking at how well they work and if they're fair to everyone.

4. Domain Awareness System in New York City

What Is It?

The Domain Awareness System (DAS) is a big network of:

• Sensors
• Databases
• Devices
• Software
• Infrastructure

It gives real-time info to police officers in New York City on their phones and computers.

How It's Used

At first, DAS was made to help stop terrorism. Now, it's used for all police work:

• Every police station in New York City has it
• Every NYPD officer has it on their phone

This means all police can get important info quickly.

Does It Work?

DAS helps NYPD work better:

What It Does	How It Helps
Shares Info	Police can share info better than anywhere else
Quick Analysis	Gives police info they can use right away
Uses Police Know-How	Mixes what police know with Microsoft tech

Problems

Even though DAS is helpful, it has some issues:

1. Privacy: People worry about being watched too much
2. Making It Fit: Other cities need to change DAS to work for them
3. Hard to Keep Up: It's tough to keep such a big system working well

Even with these problems, DAS is a big step forward in using computers to help police work.

sbb-itb-ea3f94f

5. Cloud Walk in China

What Is It?

Cloud Walk is a Chinese company that made a smart computer system to guess who might do crimes. This system uses:

• Face scanning
• Watching how people walk
• Looking at what people do
• Police records

The computer looks at how people move, what they buy, and what they do to guess if they might do something bad.

Where It's Used

Cloud Walk's system is used all over China:

• It's connected to police in more than 50 cities and areas
• It watches public places and follows people
• It tells police when it sees something odd

The system can spot things like someone walking back and forth in one place, which might mean they're trying to steal.

Does It Work?

We don't know exactly how well it works, but here's what it tries to do:

Goal	How
Stop crimes	Find people who might do bad things
Help police work faster	Tell police right away if something looks wrong
Watch more places	Follow people as they move around

Because it can look at lots of things at once and use police records, it might be very helpful for police.

Problems

Even though the system is very smart, it has some big problems:

1. People worry about privacy: It watches people all the time, which many don't like.
2. It might not be fair: Guessing who might do crimes based on how they act could be wrong.
3. It needs good information: If the computer gets bad info, it might make mistakes.
4. Other countries don't like it: Many people think it's not okay to watch everyone so much, especially some groups.

Like other systems that watch people, Cloud Walk in China makes people wonder if it's okay to watch everyone to try to stop crimes.

6. HART in Durham, UK

What is HART?

HART (Harm Assessment Risk Tool) is a computer program used by police in Durham, UK. It looks at lots of information to guess if someone might do a crime again.

How It Works

HART started in 2016. It looks at:

• Past crimes
• Age
• Gender
• Where people live

It puts people into three groups:

Risk	Meaning	Can Join Checkpoint?
Low	Not likely to do crime again	Yes
Medium	Might do crime again	Yes
High	Very likely to do crime again	No

People in the medium group can join a program called Checkpoint to help them stay out of trouble.

Did It Help?

From 2016 to 2021, HART looked at:

• 12,200 people
• 22,265 times

The computer could look at more info than people can. But we don't know exactly how much it helped.

Problems

HART had some issues:

1. Needed a lot of work to keep it running well
2. People worried it might not be fair
3. It was hard to make sure it was used the right way

Because of these problems, Durham Police stopped using HART in September 2020.

What Now?

Even though they stopped using HART, Durham Police still want to find good ways to help officers do their job well and fairly.

7. Cape Town's Gunshot Detection and Drone System

What They Use

Cape Town uses a mix of tech to fight crime:

1. ShotSpotter: Finds where gunshots come from
2. Drones: Flying cameras that watch from above
3. CCTV: 6,500 cameras around the city

How They Set It Up

They started using this tech bit by bit:

• December 2022: Put ShotSpotter in Hanover Park
• February 2023: Added it to Manenberg
• Recently: Put it in Lavender Hill too

They're also trying out drones in Hout Bay with help from local groups.

Does It Work?

The new system seems to help:

What They Found	How Many
Guns taken	More than 35
Bullets found	More than 400
People arrested	More than 50
Gunshots heard (Dec 2022)	68 times
Total shots fired (Dec 2022)	224

This shows the system helps find where crimes happen and catch people who do them.

Problems They Face

Even though it helps, there are some issues:

1. Need more police: They have to get to places fast when the system finds something
2. Talking to people: They need to make sure people are okay with the tech watching
3. Lots of info: The system makes a lot of data they need to look at
4. Working together: Different police groups need to use the tech together

Even with these problems, other cities might want to try what Cape Town is doing to stop crime.

Comparing the 7 Examples

Let's look at how these seven AI crime mapping tools stack up:

Example	What It Uses	How It's Used	Does It Work?	Problems
PredPol in Los Angeles	Smart computer program	Guesses where crimes might happen	Some say yes, some say no	Might focus too much on certain areas
ShotSpotter in Chicago	Sound sensors	Finds where gunshots come from	Often wrong about gunshots	Costs a lot, might not be fair
Hikvision AI Cameras in South Africa	Smart cameras	Watches for crimes	Helps police respond faster	People worry about privacy
Domain Awareness System in New York City	Big network of tools	Gives police real-time info	Helps police share info better	Hard to keep working well
Cloud Walk in China	Face scanning, behavior watching	Guesses who might do crimes	Not sure	Big privacy worries
HART in Durham, UK	Computer program	Guesses if someone might do crime again	Stopped using it	Wasn't fair, hard to use right
Cape Town's Gunshot Detection and Drone System	Sound sensors, flying cameras	Finds gunshots and watches from above	Helped catch some criminals	Need more police to respond quickly

When we look at these examples, we can see:

1. Different Tools: Each place uses different tech to fight crime.
2. How They're Used: Some cities use these tools everywhere, others only in some areas.
3. Do They Work?: It's hard to tell. Some say yes, some say no.
4. Common Problems:
   • Might not be fair to everyone
   • People worry about privacy
   • Police need to respond quickly
   • Lots of information to handle
5. Not Enough Info: We don't know much about some of these tools.

These AI crime tools might help keep people safe, but cities need to think carefully about using them. They need to make sure the tools are fair and don't invade privacy too much.

Ethical Issues and Problems

AI crime mapping tools raise big concerns that need careful thinking:

Privacy Worries

Systems like New York City's Domain Awareness System watch people a lot. They:

• Collect personal info
• Use face scanning
• Track where people go

Often, they do this without asking people if it's okay.

Unfair Treatment

Some AI tools, like PredPol, might treat some groups worse than others:

Problem	Result
Old crime data shows racism	Police focus too much on some areas
More police in certain spots	More arrests in those places
Secret computer programs	Hard to check if they're fair

Keeping Data Safe

AI tools collect lots of info. This can be risky:

• Bad people might steal the data
• The data might be used for the wrong things
• It's not clear how long police keep the data

People Don't Trust Police

When police use AI without telling people, it can make things worse. People might not trust the police, especially in areas where police have caused problems before.

Guessing Who Might Do Crimes

Some AI tries to guess who might do crimes. This brings up hard questions:

1. Police might stop people before they do anything wrong
2. People might be labeled as "dangerous" unfairly
3. Judges might make choices based on what the computer says

To fix these problems, police and lawmakers need to:

• Check AI tools often to make sure they're fair
• Tell people how they use AI
• Keep people's info safe
• Teach police how to use AI the right way

What's Next for AI Crime Mapping

AI crime mapping is changing fast. Here's what we might see soon:

Better Computer Learning

Computers will get smarter at finding crime patterns:

• Seeing Patterns: AI will spot tricky crime connections that people might miss.
• Guessing Future Crimes: AI will make better guesses about where crimes might happen.
• Reading Reports: AI will understand police reports and social media posts better.

New Tech Working Together

AI crime mapping will work with other new tech:

New Tech	How It Helps
Special Glasses	Show 3D views of crime scenes
Virtual Reality	Walk through crime scenes in a computer
Smart City Sensors	Get info right away
Fast Internet	Send info quickly to police

Making AI Fair

People want AI to be fair to everyone:

1. Checking for Mistakes: New ways to make sure AI doesn't treat some people unfairly.

2. Telling People How It Works: Police will explain how they use AI.

3. Asking People What They Think: Communities will help decide how to use AI.

Using Data to Make Rules

AI will help make better rules about crime:

• Stopping Crimes: Use what AI learns to make better plans to stop crime.
• Putting Police Where Needed: AI will help decide where to send police.
• Checking What Works: AI will show which police plans work best.

Wrap-up

AI crime mapping helps police work better. It looks at lots of info to:

• Find crime patterns
• Guess where crimes might happen
• Help police use their time wisely

But using AI for police work can cause problems:

Good Things	Problems
Helps stop crimes	Might not be fair to everyone
Makes police work faster	People worry about privacy
Finds hidden crime patterns	Computers might make mistakes

To use AI well, police need to:

• Make sure it's fair
• Tell people how they use it
• Keep people's info safe
• Learn how to use it right

What's next for AI crime mapping:

1. Smarter computers
2. Working with new tech like special glasses and city sensors
3. Making AI more fair
4. Using what AI learns to make better rules

Related posts

• AI Predictive Policing: Technologies, Data Sources, Fairness
• AI Predictive Policing Accuracy: 2024 Analysis
• Machine Learning Models for Crime Hotspot Mapping
• AI in Policing: Impact on Workforce & Future Trends