Should Police Be Able to Predict Crime? Professor Releases Roadmap for AI and Policing Ethics Research
Cal Poly philosophy Professor Ryan Jenkins and University of Florida professor Duncan Purves have released a 30-page report, “Artificial Intelligence Ethics and Predictive Policing: A Roadmap for Research,” funded by the U.S. National Science Foundation.
Considering that the way forward in technology ethics and policy is often unclear at the early stages, the report maps key interdisciplinary and entangled issues to guide policymakers, police and community members and to scaffold research over the coming years.
Predictive policing uses artificial intelligence to forecast future criminal behavior based on historical data. This technology is already used in more than 60 police U.S. departments.
This and similar technologies — such as facial recognition and algorithmic recidivism “risk scores” — have come under intense scrutiny in the wake of widespread concerns about bias in the computer algorithms that governments rely upon to decide the fate of citizens. This ambitious and wide-ranging report, available online here, illuminates questions that weave throughout the interplay between law enforcement, technology development, and broader society.
“These are urgent and fundamental questions that we’re advancing,” said Jenkins, the leader of the project at Cal Poly. “What's the proper role of police in society? How can technology be crafted from the beginning to be democratic and accountable, rather than marginalizing? We want to offer a structure as this research area continues to blossom, to point out the connections and nurture interdisciplinary collaboration.”
As experts in the ethics of emerging technology, Jenkins and Purves drew upon empirical use cases and a diverse, international community of experts on technology, policing and law.
Overarching questions concerning bias, the obligations of technology companies, and society’s fundamental conception of what crime is open the report and color its subsequent insights.
“We were surprised at how under-studied some of these questions were,” said Purves, co-leader of the project at the University of Florida. “But we also know that the resources exist within philosophy and adjacent fields to clarify the landscape and the strengths and weaknesses of different positions concerning the role of police and their use of technology.”
Through intensive research, the insights of the report aim to provide a base for key stakeholder deliberation and community engagement. Even though “predictive policing” as a term has fallen out of favor, the same urgent questions haunt the development of its successor: “data-driven policing.”
This underscores the need for continued interdisciplinary conversations and attention on putting ethical theories into practice throughout the entire lifecycle of technology development and deployment.
This report encompasses the preliminary first-year findings of a three-year National Science Foundation research award and offers a guide for further investigation by highlighting theoretical and concrete ethical issues, as well as empirical claims that merit further study. It is organized by audience, targeting designers and developers, police departments and police officers, and policymakers in turn.
More information online: www.aipolicing.org