Scribd
Upload
31 views

AI Individual Slides

PredPol is an AI system used by the LAPD to predict where crimes were likely to occur based on past crime data. It produced heat maps highlighting high-risk areas to assign extra police surveillance. Studies found property crime rates reduced 12% in areas where PredPol was deployed. However, critics argue it can perpetuate discrimination by over-policing minority neighborhoods and that its predictions are difficult to interpret, potentially leading to overuse of police force. Governments must balance using predictive policing AI to better police society without losing the public's trust by invading citizens' privacy rights.

Uploaded by

trex
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
31 views

AI Individual Slides

PredPol is an AI system used by the LAPD to predict where crimes were likely to occur based on past crime data. It produced heat maps highlighting high-risk areas to assign extra police surveillance. Studies found property crime rates reduced 12% in areas where PredPol was deployed. However, critics argue it can perpetuate discrimination by over-policing minority neighborhoods and that its predictions are difficult to interpret, potentially leading to overuse of police force. Governments must balance using predictive policing AI to better police society without losing the public's trust by invading citizens' privacy rights.

Uploaded by

trex
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 18

Case Study:

PredPol, as used
by the LAPD
Introduction, Benefits, Limitations, Analysis
01
Introduction
What PredPol is, and what it does
What is PredPol?
AI system used by the LAPD to predict
where crimes were likely to happen. It
claimed to make predictions by only using
past crime locations, times and dates.
Heat maps
PredPol would produce heat maps
highlighting where crime was most
likely to happen. Highlighted areas
would be assigned extra surveillance
by law enforcement

Source: PredPol (2020)


Optimized Patrols
PredPol also helped optimize patrol
routes and monitor officers’
movements for more effective
policing

Source: PredPol (2020)


Heat maps
PredPol would produce heat maps
highlighting where crime was most
likely to happen. Highlighted areas
would be assigned extra surveillance
by law enforcement

Source: PredPol (2020)


02 Benefits
Effectiveness of PredPol
Reduction in property crime rates
PredPol not
Deployed

PredPol
Deployed

Mohler (2015) found that property crime rates in areas where PredPol
was deployed were reduced by 12%!
Manpower Allocation

PredPol increased manpower allocation efficiency!

● Police now focus on high-risk areas

● Able to dynamically change patrol routes

● GPS tracking holds officers accountable on patrol

Quicker response times to emergencies benefited the


community!
Finances
PredPol lessened the financial strain on police departments!

● Reduces manpower requirement

● More responsible and effective usage of the public’s tax


money!
But,
was PredPol all good?
03
Limitations
You can enter a subtitle here if you need it
Perpetuates Discrimination

Past data (which PredPol relies on) is often biased against minority communities.
This leads to over policing of minorities which can create a self fulfilling prophecy -
greater police presence resulting in a greater number of crimes caught in the area,
reconfirming the AI’s predictions.
Inability to interpret data
According to Moses & Chan (2018):

● Predictions made by PredPol are hard to interpret

● Poor interpretation leads to over policing and biased


enforcement

● Models require constant updates by those with on the


ground knowledge

Also, the AI model does not instruct on how to interact with crime scenarios! Officers may treat all
cases as dangerous scenarios, leading to overt use of force!
Source: Straits Times (2021)

Privacy
Implications
Data farming for the AI may
infringe upon the rights of
private citizens, most of whom
are law abiding

When the police exploited a loophole in the


law to use covid tracing data for
investigations, many Singaporeans viewed
it as an invasion of privacy!
04
Analysis
Balancing Act

Using AI for Maintaining trust


predictive policing with the public

As using AI has serious implications on the privacy rights of individuals,


governments must be delicate in how they use AI technology in policing,
so that they can better police society without losing the public’s trust
Works cited
Bennett Moses, L., & Chan, J. (2018). Algorithmic prediction in policing: Assumptions,
evaluation, and accountability. Policing and Society, 28(7), 806–822.

Chee, K. (2021, February 2). Vivian Balakrishnan says he 'deeply regrets' mistake on
Tracetogether Data. The Straits Times. Retrieved April 15, 2023, from
https://www.straitstimes.com/singapore/vivian-balakrishnan-says-he-deeply-regrets-mist
ake-on-tracetogether-data-first-realised-it

Mohler, G. O., Short, M. B., Malinowski, S., Johnson, M., Tita, G. E., Bertozzi, A. L., &
Brantingham, P. J. (2015). Randomized controlled field trials of predictive policing.
Journal of the American Statistical Association, 110(512), 1399–1411.

PredPol. (2023). How PredPol Works | Predictive Policing. PredPol.


https://www.predpol.com/how-predictive-policing-works/

You might also like