• Corpus ID: 3318760

Runaway Feedback Loops in Predictive Policing

  title={Runaway Feedback Loops in Predictive Policing},
  author={Daniel L. Ensign and Sorelle A. Friedler and Scott Neville and Carlos Eduardo Scheidegger and Suresh Venkatasubramanian},
Predictive policing systems are increasingly used to determine how to allocate police across a city in order to best prevent crime. [] Key MethodIn response, we develop a mathematical model of predictive policing that proves why this feedback loop occurs, show empirically that this model exhibits such problems, and demonstrate how to change the inputs to a predictive policing system (in a black-box manner) so the runaway feedback loop does not occur, allowing the true crime rate to be learned. Our results…

Figures from this paper

A Data-driven analysis of the interplay between Criminological theory and predictive policing algorithms
The study found that multidisciplinary analysis of such systems is vital for uncovering these issues and shows that any study of equitable AI should involve a systematic and holistic analysis of their design rationalities.
The effect of differential victim crime reporting on predictive policing systems
It is demonstrated how differential victim crime reporting rates across geographical areas can lead to outcome disparities in common crime hot spot prediction models, which may lead to misallocations both in the form of over-policing and under-Policing.
A Penalized Likelihood Method for Balancing Accuracy and Fairness in Predictive Policing
It is shown that fairness can be introduced into point process models of crime so that patrol levels proportionally match demographics, though at a cost of reduced accuracy of the algorithms.
Predictive Policing and Crime Control in The United States of America and Europe: Trends in a Decade of Research and the Future of Predictive Policing
Examining predictive policing over the last decade reveals that there is consensus by the majority of the researchers on the importance of predictive algorithms on the policing landscape, and there are several impediments that have militated against the effectiveness of predictive policing.
Utility of predictive policing: Evaluating a Spatial-Temporal Forecasting Model in an operational deployment
The evaluation indicated the pilot project had a quantifiable effect in 4 out of the 6 months, with inconclusive findings for the remaining, and was designed to contribute to the published literature on predictive policing and to offer guidance for police services considering adoption of this technology, by providing a template that could be used to assess the strategy’s effectiveness.
Dirty Data, Bad Predictions: How Civil Rights Violations Impact Police Data, Predictive Policing Systems, and Justice
Law enforcement agencies are increasingly using predictive policing systems to forecast criminal activity and allocate police resources. Yet in numerous jurisdictions, these systems are built on data
Fairness in Algorithmic Policing
  • D. Purves
  • Law
    Journal of the American Philosophical Association
  • 2022
Predictive policing, the practice of using of algorithmic systems to forecast crime, is heralded by police departments as the new frontier of crime analysis. At the same time, it is opposed by
A review of predictive policing from the perspective of fairness
This work evaluates the existing publications in the field of fairness in machine learning and predictive policing to arrive at a set of standards for fair predictive policing and urges that the growing literature on fairness in ML be brought into conversation with the legal and social science concerns being raised about predictive policing.
The End of Intuition-Based High-Crime Areas
In 2000, the Supreme Court held in Illinois v. Wardlow that a suspect’s presence in a “high-crime area” is relevant in determining whether an officer has reasonable suspicion to conduct an
Predictive Policing and the Ethics of Preemption
This chapter examines the increasing use of predictive analytics technologies by police. Such tools analyze data to identify patterns and make predictions about crime. Champions of predictive


Randomized Controlled Field Trials of Predictive Policing
The concentration of police resources in stable crime hotspots has proven effective in reducing crime, but the extent to which police can disrupt dynamically changing crime hotspots is unknown.
Predictive Policing: The Role of Crime Forecasting in Law Enforcement Operations
This guide assesses some of the most promising technical tools and tactical approaches for acting on predictions in an effective way for preventing crime, solving past crimes, or identifying potential offenders and victims.
To predict and serve?
Predictive policing systems are used increasingly by law enforcement to try to prevent crime before it occurs. But what happens when these systems are trained using biased data? Kristian Lum and
Statistical Procedures for Forecasting Criminal Behavior
This paper addresses the apparent contradiction between claims that for criminal justice applications, forecasting accuracy is about the same and procedures such as machine learning that proceed adaptively from the data will improve forecasting accuracy.
Rawlsian Fairness for Machine Learning
This work studies a technical definition of fairness modeled after John Rawls' notion of "fair equality of opportunity", and gives an algorithm that satisfies this fairness constraint, while still being able to learn at a rate comparable to (but necessarily worse than) that of the best algorithms absent a fairness constraint.
Fairness in Learning: Classic and Contextual Bandits
A tight connection between fairness and the KWIK (Knows What It Knows) learning model is proved: a provably fair algorithm for the linear contextual bandit problem with a polynomial dependence on the dimension, and a worst-case exponential gap in regret between fair and non-fair learning algorithms.
Fairness Incentives for Myopic Agents
This work investigates whether it is possible to design inexpensive subsidy or payment schemes for a principal to motivate myopic agents to play fairly in all or almost all rounds, and shows both positive and negative results in the classic and linear bandit settings by upper and lower bounding the cost of fair subsidy schemes.
Predicting How People Play Games: Reinforcement Learning in Experimental Games with Unique, Mixed Strategy Equilibria
The authors examine learning in all experiments they could locate involving one hundred periods or more of games with a unique equilibrium in mixed strategies, and in a new experiment. They study
Fair Learning in Markovian Environments
This work defines a fairness constraint requiring that an algorithm never prefers one action over another if the long-term (discounted) reward of choosing the latter action is higher, thus establishing an exponential gap between exact and approximate fairness.
SAMHSA provides links to other Internet sites as a service to its users and is not responsible for the availability or content of these external sites. SAMHSA, its employees, and contractors do not