Predictive Policing and Artificial Intelligence

Cover
Half Title
Series Information
Title Page
Copyright Page
Dedication
Table of contents
Illustrations
Foreword
Contributors
Introduction
	Two extremes
	A challenging environment for police forces
	What is predictive policing?
	The ‘predictive’ part
	Environmental criminology and crime science
	Artificial intelligence
	The ‘policing’ part
	AI in policing
	Structure of book
	Part I: Bias and big data
	Part II: Police accountability and human rights
	References
Part I Bias and Big Data
	Chapter 1 The future of AI in policing: Exploring the sociotechnical imaginaries
		Introduction
		1. Sociotechnical imaginaries
		2. The benefits and risks of AI for society
			Definition of AI
			The benefits of AI
			The risks of AI
				Technical limitations
				Data-driven biases
				Trust
		3. Using AI in policing
			A utopian view
			A social science view
				Assumptions
				Evaluation
				Accountability
			A data science view
			A civil rights community view
		Conclusion
		Notes
		References
	Chapter 2 Predictive policing through risk assessment
		Introduction
		Projected benefits of predictive policing with individual risk
		Examples of predictive policing tools with individual risk
		Contentious issues with individual risk prediction
		Entry points for biases in predictive policing algorithms
			Label bias
			Feature selection
			Sample bias
			Feedback loop
		Future prospects for predictive policing
		Conclusions
		References
	Chapter 3 Policing, AI and choice architecture
		Introduction
		The ubiquity of choice architecture
		Policing and choice architecture
		AI and choice architecture
		AI as a product of choice architects
		AI technologies as choice architects
		Choice architects within police organisations
		Conclusion
		References
	Chapter 4 What big data in health care can teach us about predictive policing
		Introduction
		Part I
		Predictive analytics in policing and health care
			Predictive policing
			Health care
		Part II
		The professions in dialogue
			Practitioners
				Role disruption
				Automation bias and discretion
			Policymakers
				The duty of explanation
				Transparency and trade secrets
				Scarcity and the inevitability of distributional choices
			The polity
				Bias and equality
				Privacy
		Conclusion
		Acknowledgement
		Notes
		References
	Chapter 5 Artificial intelligence and online extremism: Challenges and opportunities
		Introduction
		An overview of existing approaches
			Analysis
		Detection
		Prediction
		Challenges
		Defining radicalisation
		Data collection, verification and publication
		Noisy data (false positives)
		Biases
		Incompleteness
		Heterogeneity (variety of content)
		Irreproducibility
		Research methodologies
			Lack of comparison against a control group
			Lack of comparison across approaches
			Lack of cooperation across research fields
			Adaptation of extremist groups
		Ethics and conflicts in legislation
		Opportunities
			Collaboration across research disciplines and organisations
			Creation of reliable datasets to study radicalisation
			Comparative studies
			Contextual adaptation of technological solutions
			Better integration of humans and technology
			Ethical vigilance
		Conclusions
		References
	Chapter 6 Predictive policing and criminal law
		Introduction
		Part I: Crime prevention and law enforcement
			A. Rational offenders and the expected benefits and costs of crime
			B. Punishment-focused deterrence
			C. Police-focused deterrence
			D. Long-term and short-term deterrence
			E. Real-time policing and enforcing the criminal law
		Part II: Machine predictions and policing
			A. Machine learning: a brief overview
			B. Place-based predictions
			C. Person-based predictive policing
			D. Real-time situational awareness technologies
		Part III: Real-time policing and crime prevention
			A. Inchoate and corollary crimes
			B. Precommitment and credible law enforcement policies
			C. Real-time policing, salient signals and deterrence
				Myopic offenders
				Self-control problems and time-inconsistent misconduct
				Perceptual deterrence and ‘erroneous crimes’
				Learning from crime and serial offenders
			D. Real-time intervention
				The projection bias and hot-state crimes
				Risky crimes
		Part IV: The social costs of relying on machine predictions in policing
			A. Fairness and accuracy
			B. Machine predictions and indirect, non-transparent deterrence
			C. Switching to more serious crimes under a proactive predictive policing regime
			D. Machine predictions and police judgements
			E. The costs of proactive deterrence policies
		Conclusion
		References
Part II Police accountability and human rights
	Chapter 7 Accountability and indeterminacy in predictive policing
		Introduction
		Police, accountability, transparency and reform
		Three accountabilities
		Algorithmic indeterminacy
		Towards a police accountability frame for the age of AI
		References
	Chapter 8 Machine learning predictive algorithms and the policing of future crimes: Governance and oversight
		Introduction
		Functions of the police in England and Wales under the common law
		‘Austerity AI’ and the problem of prioritisation
		‘In accordance with law’
		Discretion in police decision-making
		Impact on rights
		Safeguards, governance and oversight
		Conclusion
		Notes
		References
	Chapter 9 ‘Algorithmic impropriety’ in UK policing contexts: A developing narrative?
		Introduction
		Algorithms in the UK public sector
		The Gangs Matrix case study
		The West Midlands case study
		Legal points on algorithmic or predictive policing tools
			Data scope issues
			Process issues
			Issues of human rights impacts
		Conclusions
		References
	Chapter 10 Big data policing: Governing the machines?
		Introduction
		The problem of governance
		The problem of privacy
		The problem of bias
		Conclusion
		References
	Chapter 11 Decision-making: Using technology to enhance learning in police officers
		Introduction
		Artificial intelligence
		The demands of modern-day policing
		Training and upskilling the next generation of officers
		Contextualising learning
		Developing a personalised reflective learning environment for policing using technology
		Policing exemplar #1
		Created immersive learning environments
		Policing exemplar #2
		Responsive immersive learning environment: application of the decision-making framework
		Policing exemplar #3
		Summary and future directions
		References
	Conclusion
		References
Index