Cspan. Org, or listen on the free cspan radio app. Susan sarah brayne, your new book seems like it is welltimed for a National Debate on policing, but you tell readers youve been working on the project about a decade. How did you get started in this interest in big data and the police . Sarah when i was a phd student at princeton i think it was back in 2012, this is when there was the start of enthusiasm over big data was happening. People were saying big data is transforming everything from finance to sports to journalism, marketing, insurance, education. But no one was yet working on how big data would or would not transform the criminal Justice System. Id had a longstanding interest in the criminal Justice System and i started to ask, how are the police, courts, corrections, leveraging things like predictive algorithms and how is it changing daily operations . I quickly realized there was not actually ironically very good data on police use of big data, and thats when i decided to pursue an ethnographic study on that question. Susan we will have lots of time to explore the details, but what is the conclusion you came to after you spent this amount of time investigating the topic . Sarah the conclusion is basically that instead of thinking about data as some sort of objective or fundamentally unbiased tool, i think it is better to think about data as situated in social systems and organizations. Its like a form of capital, some people have more or less of, it can be used for different purposes and it has an increasing amount of value in organizations. It can be used to achieve different types of organizational goals. The book disabuse us folks of the notion that data is an objective and unbiased force that can fix all of the problems that have existed for decades if not centuries. Susan in effect, it is a cautionary tale for people who are looking at the use of data to solve some of the policing problems the country is having. Sarah correct. Susan you write that big data and the criminal Justice System motivates my work. First, define what data is. What is big data . You talk about volume, velocity, and variety. Sarah typically it refers to volume, meaning there is a lot of different pieces of data. Second, it has to do with variety. The data comes from a range of disparate institutional sources. The police have always collected their own crime data, for example, but increasingly the police are using information from a range of institutional sources, like private data brokers, social media data. That is the variety part of it. Velocity has to do with processing and sometimes Storage Capacity as well. That means you can run analysis on this best of these vast troves of data that previously wouldve taken days or weeks, but with increasing computing power, we can run them almost instantaneously and often remotely as well. Susan why do you come to the conclusion that the u. S. Criminal Justice System is overgrown . Sarah sort of no matter how you measure it, the u. S. Criminal Justice System is unprecedented in size and scope, both in International Comparative perspective, looking at p nations, but also historically. At p nations, but historically. Its been on this ride since the 1970s. Also if we look at federal expenditures directed toward local Law Enforcement agencies or the number of cops on the street, the United States is without peer in terms of the size and scope of the criminal Justice System. Sometimes scholars talk about the phenomenon as mass incarceration, chest to do with not just the reach of who is behind bars but the families, communities, neighborhoods that are heavily surveilled and policed as well. Susan one statistic you give is that 70 million americans have a record on file with criminal justice agencies, that is out of a population of 330 million americans. Is that statistic a surprising one compared to other western nations . Sarah it is definitely very surprising and also surprising to the audiences i present to as well. Ive had many folks in the audience try to correct me on that, saying you must have an extra zero in there and that kind of thing. There is a tendency sometimes to think about involvement in the criminal Justice System as this sort of unusual or uncommon experience, whereas increasingly, particularly for certain demographic groups such as those with lower levels of Educational Attainment or racial minorities, involvement in the criminal Justice System is a modal life experience. I think it is a surprising statistic, especially when you look in comparison to other countries with similar crime rates. Susan to write this book and do this research, you spent a lot of time in the field. Tell me about that experience. Sarah basically, as i mentioned, theres not very much data on police use of big data, so what that called for was what we call ethnographic fieldwork, which is talking to people, watching people as they do they go through their daily lives. I quickly realized i needed to get access to a Police Department and i selected the lapd not because it is an average Police Department, it is not, it is large and wellfunded. But i selected it for those reasons, it is on the forefront of police use of Data Analytics and i thought it could broadcast some trends that could shape a smaller Police Departments in coming years. My fieldwork involved observations and interviews with sworn officers and civilian employees in different areas and specialized divisions of lapd. Things like the Information Technology division, records and identification, robbery homicide, fugitive warrants. I also went on ride alongs and did some observations at the joint Regional Intelligence Center in southern california. For those who are not familiar Fusion Centers are basically , these multidisciplinary, multi agency surveillance organizations funded by the federal government, largely in the wake of 9 11, which was seen as an information sharing failure for the intelligence community. I also did interviews with folks who work at Key Technology companies who designed some of the Analytic Software the lapd uses. Those include things like palantir. Susan was your presence in the Police Department welcomed for the most part . Sarah thats an interesting question. I would say i was able to obtain a great degree of access, unprecedented in some ways. But there was a lot of variation in how i was received, i think. This is one of the things that Ethnographic Research exposes. Theres a lot of heterogeneity in Police Department. The people who were very proud of how the department was using data, they consider themselves on the vanguard and leaders in the field, and they were excited to have me there and show me the new technological tools they had. But at the same time, historically police have not been particularly receptive to researchers or journalists spending time with them. There was a certain garden this or resistance and other spaces where they were concerned i was going to do a hit piece on them, that i was going to come in and do a gotcha story and leave. I think the resistance to my presence declined the longer i spent in the field, because i was able to build some of those relationships and share my findings with the police themselves. There was also a lot of confusion about who i was, to be honest, in the field, because Police Department are very hierarchical organizations. If someone higher up would vouch for my presence, they would say the sergeant would say it is fine, and then the patrol officers would find that to be tolerable. But they often confused who i was. They would say, instead of saying i was a sociologist from princeton, they would say she is a psychologist from harvard or a senior from stanford and that kind of thing. There was definitely confusion about my presence in the field as well, particularly on ride alongs. Susan it was important to select lapd, but how many Police Departments around the United States are actually using software and big Data Collection . Sarah unfortunately, i dont have the answer to that question. I think that is a fundamental problem when we are talking about, we are having nationwide conversations about what policing should and should not look like. There isnt any centralized data on this. There are some national surveys. They are pretty out of date, it is voluntary compliance, and a lot of the time, the budgetary allocations to these kind of advanced technologies are pretty ambiguous. You cant necessarily always tell from a Police Budget amount of money is spent on this kind of thing. I think it would probably fall somewhere in the 70 range from what i have read, with the departments not using these kind of tools largely being small Police Department. The United States, another thing that is unique about the criminal Justice System, we have this system of policing where there are a lot of Police Departments that are extremely small, under 20 officers, and they lack the training, funding, infrastructure for these technologies. Susan one other question about your preparation and research, you founded something called the texas Prison Education initiative. What is it and how did it inform your work on this topic . Sarah it is basically just a group of volunteer professors and graduate students, postdocs, who teach classes in prison. My was in graduate school, i started volunteer teaching and state prisons in new jersey, just volunteer Teaching College classes. When i moved to texas, which has one of the largest prison populations in the country, i figured the university here would have the same sort of Prison Ed Program to volunteer with. For me, it is important because it is a skill set around being an educator, and also i do research on the criminal Justice System. It seems unwise to have no exposure to folks whose lives are on the line in the system. There wasnt a Prison Education program at ut at the time. Along with a fellow graduate student, we founded the texas Prison Education initiative and now we offer College Credit classes to about hundred 50 folks. 150 folks. Everything from sociology to physics, english, math. The courses are for credit in available without cost to the students. Susan how did that work inform your conclusions on pig data and on big data and policing . Sarah one of the things that came up that is informing my future research, is a lot of the time there is a lack of knowledge or information among the folks incarcerated about the particulars of their case. Particularly the police investigation. I would have students say things to me like my lawyer had no idea, my defense attorney had no idea was under suspicion. They did not know why the cops were sitting outside my house that day. One of the things about big data policing is it is largely invisible. When we have a lot of cops on the street, that is a very visible Police Presence you can understand and feel. Big data policing can be invisible and it is hard to put your finger on. What actually is going on in Police Investigative procedures is largely black ops. I think this exacerbates preexisting inequalities in the system. Susan you mentioned there was a big uptick following 9 11, thats also when people began to see Police Department around the country were bringing surplus military equipment into their use. That decision is being reexamined both at the state and national level. You write the creep of military software into Police Operations is largely overlooked. Why is that . Sarah i think it is precisely because it is so invisible. As you said, theres been increased attention about the militarization of policing, how Police Departments are getting surplus military kit, small, local police permits have tanks and that sort of thing. I think military software is this largely untold part of the story, where a lot of different techniques and platforms and algorithms that were initially designed for military applications are being imported into local policing. I think the reason we dont know much about it is because it is largely invisible. You can see a tank rolling down the street at the pumpkin festival in New Hampshire but you cant necessarily see they are using predictive algorithms. Susan Police Department have used data, even if it was manually collected, for a long time. Is the issue here that rather than reactive policing, solving a crime that has happened, it is the use of data to predict where crime might happen and how that is used and whether it is used equally . Is that the question for society . Sarah i think that is one of the nubs of the question, weve had this shift from reactive to proactive policing. Another key transformation that is also occurring, as you said, the police have been long collecting their own data, but that is on people they have contact with. One is happening in the digital age is the police are increasingly collecting information on all of these folks who have no direct criminal justice contact. Part of that has to do with the variety component of the three vs of big data, they are purchasing information from private companies, they are using tools like automatic license plate readers. You dont have to get pulled over for your data to be put in their system. That kind of information is also being used in the predictive modeling, for example. Susan tell me a little more about the companies you mentioned. The first is palantir technologies. What does it do for Police Departments . Sarah they actually just went public in september of this year, so there is a little more Information Available on them now than before, but when i started this, very few people had heard of them. Essentially they have a platform that is called Palantir Gotham where Law Enforcement agencies and other clients are able to aggregate all of these disparate sources of information and visualize them in certain ways. For example, the police might have their own crime data, but you can build out these networks where somebody at the center of the network might have direct Police Contact and be stopped by the police, for example, but there is a secondary Surveillance Network lets say i was stopped by the lapd and you never have. I called you on a phone and now your phone number is associated with it. Or i was parked outside of your house. Or it has information on who i am dating or who my siblings are or where i work. You have radiating out this secondary Surveillance Network of individuals who dont have any Law Enforcement contact. Thats one way to use it, to visualize different sources of data. It is also used for investigations. For example, there was an instance in which there was some copper wire theft going on in the city, and they were able to basically draw a radius and palantir had the mapping function, and they were able to list the vehicles in the radius between midnight and 5 00 a. M. When they knew the wire theft had occurred, and narrow it down to three cars that mightve been involved in the crime. It is this very powerful Data Integration and surveillance tool that can yield insights that might have previously taken hundreds of detectives on a case, or weeks or months, and of this shoe leather policing, and you can use it to narrow that down. Susan people listening might say that sounds like a good thing, use less police resources. Wheres the problem . Sarah i think the issue lies in the sense that, in order for the system to work perfectly, we have to assume infallibility of the state. That the state never makes mistakes, and the inference one draws about why you are in a place at a given time are unbiased and without error. And everybody entering the data there is still a human side of Data Collection, its not all automated but anybody who enters the information will do so without bias or error or prejudice come and that is just not borne out in ethnographic fieldwork. We have tons of information that suggests the data inputted in these police systems, for example, is as much a function of Law Enforcement practices as it is actual ascending rates for we also have a lot of information that error is disproportionately distributed. There is a study out of michigan indicating that about that black folks are about seven times more likely to be wrongly convicted of murder than whites. In order to be ahead, if you think about dna databases, you have to be in the database in the first place. Part of the challenge with big data policing is