An excerpt and interview with Andrew Ferguson, author of “The Rise of Big Data Policing”
By WASHINGTON, D.C.|J.F.
“Minority Report”, a 2002 film directed by Steven Spielberg, features a squad of police officers who arrest people for murders they are predicted to commit. The film was science fiction; yet police departments around the world increasingly use predictive analytics to identify people who might become perpetrators or victims of crime. In “The Rise of Big Data Policing”, Andrew Ferguson, a former public defender and now professor at the University of the District of Columbia, discusses the promise and perils of data-driven policing.
The Economist asked him about how data and predictive analytics are changing modern policing. After his responses, you can read an excerpt from his book that shows what data-driven policing looks like on the ground.
The Economist: Police have always used data to make decisions. What makes this era different?
Andrew Ferguson: Policing has traditionally been reactive: officers respond to calls for service, and experience determines where they patrol. Big-data technology lets police become aggressively more proactive. New data sources coupled with predictive analytics now allow police to visualise crime differently, targeting individual blocks, at-risk individuals and gangs in innovative ways. New surveillance technologies let police map physical movements, digital communications and suspicious associations in ways that can reveal previously hidden patterns of criminal activity in otherwise overwhelming amounts of data. All of this information can be quite useful to law enforcement seeking to track criminal elements in society. The same technology can also be quite threatening to civil liberties and personal privacy in already over-policed communities.
More than 60 American police departments use some form of “predictive policing” to guide their day-to-day operations
The Economist:How pervasive is the use of tech in policing—how different is the day-to-day work of police officers today as opposed to 20 or 30 years ago?
Mr Ferguson: Technology is shaping where police patrol, whom they target, and how they investigate crime. More than 60 American police departments use some form of “predictive policing” to guide their day-to-day operations. In Los Angeles, this means that police follow patrols based on computer-forecast crime hot-spots. In Chicago, an algorithmically derived “heat list” ranks people at risk of becoming victims or perpetrators of gun violence. The result is that police prioritise particular places and people for additional contacts and monitoring. In addition, new surveillance technologies—including police body cameras, automated licence-plate readers, Stingray cell phone trackers and high-definition surveillance cameras—provide powerful monitoring tools. All of this technology changes how officers see the communities they patrol and the citizens they police. The technology also changes the job of policing, forcing officers to become data collectors and analysts as they act on real-time inputs and assessments.
The Economist: Does big-data policing work? Has it made people less likely to be victims of crime?
Mr Ferguson: The jury is still out on effectiveness. The scientific studies are few in number and largely inconclusive. In some cities crime rates have trended down with the introduction of new technologies, but in others there has been no significant effect. Crime rates correlate with a host of economic and environmental forces that make it difficult to demonstrate any causal connection with a specific technology. But really, the benefit of big-data policing for police departments is political. New technology gives police chiefs an answer to the age-old question asked by every politician in every community forum: “Chief, what are you doing about crime?” They now have a progressive-sounding, technologically inspired answer: “We are using a new black-box technology to predict and deter crime.” Whether it works is secondary to having a response to the otherwise unanswerable (and somewhat unfair) question that every police chief faces.
The Economist: What are the biggest potentials for abuse?
Mr Ferguson: There are several. First, data can distort policing. Officers sent to an area flagged as being at risk of violent crime may see routine encounters as more threatening, thus making them more likely to use force. Second, the growing web of surveillance threatens to chill associational freedoms, political expression and expectations of privacy by eroding public anonymity. Third, even with the best use policies in place, officers have access to vast amounts of personal information of people not suspected of any crime. Finally, without carefully chosen data inputs, long-standing racial, societal and other forms of bias will be reified in the data.
The growing web of surveillance threatens to chill associational freedoms, political expression and expectations of privacy
The Economist:How can citizens best protect themselves from such abuse?
Mr Ferguson: The time to respond to the threat of big-data policing is now. Every city should have formal written policies in place detailing the approved use of new big-data policing technologies. Every citizen should be educated about the dangers to privacy, liberty and the imbalance of power that surveillance technologies bring. Every police department should engage impacted communities about the risks and rewards of new predictive technologies with official answers to concerns about transparency, racial bias and constitutional rights. Every community should host annual “surveillance summits” where the city officials, engaged citizens and police chiefs can come together for a moment of public accountability about the use and potential misuse of new big-data technologies. Education, empowerment, and engagement are the only protections against an encroaching data-driven surveillance state.
The Violence Virus An excerpt from “The Rise of Big Data Policing: Surveillance, Race, and the Future of Law Enforcement”
A knock on an apartment door. A man gives the prognosis to a worried mother. Your son might die. He is at grave risk. Others he knows have already succumbed. An algorithm has identified the most likely to be stricken. He is one of a few hundred young men (approximately 0.048% of the city) who may die. In Chicago, Illinois, this scene has played out hundreds of times at hundreds of doors. The danger, however, is not some blood-borne pathogen. This is not a doctor giving a cancer diagnosis but a police detective giving a life diagnosis. Violence is contagious, and you are exposed. As a young man in Chicago, due to your friends, associates, and prior connection to violence, you have been predicted to be the victim or perpetrator of a shooting. Your name is on the “Strategic Suspects List,” also known as the “heat list,” and a detective is at your door with a social worker and a community representative to tell you the future is not only dark but deadly. A vaccine exists, but it means turning your life around now.
In Chicago, 1,400 young men have been identified through big data techniques as targets for the heat list. Software generates a rank-order list of potential victims and subjects with the greatest risk of violence. In New Orleans, Palantir has partnered with the mayor’s office to identify the 1% of violent crime drivers in the city. In Rochester, New York, and Los Angeles, similar techniques are being used to identify juveniles who might be involved in repeated delinquent activity. This is the promise of big data policing. What if big data techniques could predict who might be violent? What if a policing system could be redesigned to target those who are at-risk in a neighbourhood before the shooting occurs? This is the theory behind “person-based targeted policing.”
Person-based predictive policing involves the use of data to identify and investigate potential suspects or victims. Part public health approach to violence and part social network approach to risk assessment, big data can visualise how violence spreads like a virus among communities. The same data also can predict the most likely victims of violence. Police data is shaping who gets targeted and forecasting who gets shot.
While these predictive technologies are excitingly new, the concerns underlying them remain frustratingly old-fashioned. Fears of racial bias, a lack of transparency, data error and the distortions of constitutional protections offer serious challenges to the development of workable person-based predictive strategies. Yet person-based policing systems are being used now, and people are being targeted.
***
There are four main ways in which data and predictive analytics fundamentally change how police in liberal societies operate.
First, big data makes police more proactive. Traditionally, officers might react to calls for service, rely on observations made while on patrol, or respond to community complaints. With person-based predictive targeting, police can instead target suspects for surveillance or deterrence before a call comes in. For local prosecutors, this represents a significant change. As a former head of the Manhattan Criminal Strategies Unit stated, “It used to be we only went where the cases took us. Now, we can build cases around specific crime problems that communities are grappling with.”
Second, seeing violence as a public-health problem, rather than just a law-enforcement problem, lets societies rethink how best to identify and respond to criminal risk. Violence-reduction strategies in New Orleans, for instance, included social-service programmes. The idea that violence is contagious suggests that it can be prevented. If a good percentage of shootings are retaliatory, then one can design a cure that interrupts the cycle. Every time police summon people whom predictive analytics have identified as potential perpetrators or victims of violence, social-services representatives should be there, ready to offer those young men and women the opportunity to change their environment.
Third, moving from traditional policing to intelligence-led policing creates data-quality risks that need to be systematically addressed. Intelligence-driven systems work off many bits of local intelligence. Tips, crime statistics, cooperating witnesses, nicknames, and detective notes can get aggregated into a large data system. But the quality of that data is not uniform. Some tips are accurate; some are not. Some biases will generate suspicion, and some informants will just be wrong. An intelligence-driven policing or prosecution system that does not account for the varying reliability and credibility of sources—and just lumps them all together as “data”—will ultimately result in an error-filled database. Especially when these systems are used to target citizens for arrest or prosecution, the quality-control measures of black-box algorithms must be strong.
Fourth, other data-integrity concerns may arise when detectives, gang experts, or police intelligence officers control the target lists. While these professionals generally have close connections to the community and valuable knowledge of local gangs and potential targets, risk scores can be manipulated by police interested in prosecuting certain individuals. People can get added to the target lists—which are often riddled with errors—with no way to challenge or change their designation. After all, joining a gang is rarely a formal process; rumour, assumptions or suspicion can be enough to earn an elevated risk score. Worse, there is usually no easy way to get off the list, even as people’s circumstances change, time passes, and the data grows stale. The people on these lists, and most impacted by these risks, are primarily young men of colour. This reality raises serious constitutional concerns and threatens to delegitimise person-based predictive policing strategies.
