So some Risk Assessment. Are calculated. And the items variy at all of them contained history and the rest simply a few different categories, and conclude and bills, education and demographic variables and particularly gender age and include psycho social and in a corrections official and they ask things et cetera. Explicit variables and rather this was not always true and many words until explicit based but the modern instruments because theyre essentially indicators of poverty and so we can expect these instruments to have a racially impact. In my view many of the variables that go into the scores are simply inappropriate and possibly illegal factors on which to base the treatment of criminal defendant. And im not saying that those things are not correlated or predictive of recidivism. Crime is obviously correlated with poverty in our society, although we should remember that the extent of that correlation may be exaggerated by what the fact the recidivism they measure getting caught, right. So what youre really saying is that people who have these rarkt r character ristics may refle, simply, policing differences. So but even if we assume that all those characteristics actually predict crime rates, that doesnt mean that we should use them to decide how much to punish a person. When we tell a judge, sentence this person based on risk score and then we treat indicators of poverty or demographics, et cetera, as risk factors in that score, we are explicitly basically telling judges we want to literally punish people for their poverty, that is their score is determined not by what they did in the case but by who they are and their families are and how much money they have. So i have set forth a constitutional argument against the use of some of these variables, its an article, i dont have time to go through the details. I will say its not especially boundary pushing, its quite strong pushing that says, well, it comes from a case in which the state tried to revoke the probation of a probationer who lost his job, and the state tried to defend this in part by saying, look, there are all these studies that show people that people have lost their jobs and poor people are more dangerous. And the court said, look, you cant lump all poor people together and say that just because this person shares a characteristic with people who committed an elevated level of crime in the past, that this person is more dangerous, that is nothing more than punishing a person for his poverty, thats exactly how the Supreme Court described it. That is exactly what todays Risk Assessment instruments do and i think they run against that doctrine as well as what should be our moral intuitions. Its somewhat surprising that these Risk Assessment instruments have been embraced by what, otherwise, Progressive Advocates of criminal Justice Reform. Its long a prong in many of the pending risk criminal Justice Reform pieces of legislation and why is that, i think in part, people hope that it will help to reduce incarceration by allowing judges to identify low risk people who dont need to be incarcerated. My view is that, you know, that reducing incarceration is an important objective, of course, but that adopting it through measures that prevent the very people that have been the most disproportion natalie incarcerated, people who face all those from taking advantage of it is not the way to go about it. All right. I have much more to say, but im getting a big stop sign, so, ill stop. [ applause ] andrew is going to talk about predictive policing. In my own name, i would like to make one major point, when it comes to predictive policing, the color of surveillance is not white. Its not black and its not brown, its dark. Its hidden, its secret. And its hard to figure out. Predictive policing the idea that packs crime data can allow predictive hot places, hot spots or hot people, persons, to alga rhythmic number crunching. Race is usually removed from all of the algorithms. It has impacts on racial impacts. The police data comes from police and all of the implicit and explicit biases in our policing in America Today show up in that data. In this way predictably may not be any worse than traditional policing and the real question is whether it could be better. I want to sketch out the reality and give you some terms and we will discuss it in the question and answer. To give you the since of scale, its currently being used in major cities like new york, miami, santa cruz, chicago, kansas city and whole host of smaller cities. Three main types, youve got your predictive policing, ideas that you can predict, where car thefts might be, where thefts from autos might be. You have your violence, placed base Violent Crimes, gang shooting, bars an clubs and you have your person base predictions we can identify the people involved in criminal activity and target them through Public Health models or surveillance. I want to sketch out how what these look like and then well talk in question and answer what to do about them. Predictive policing, crime data, patch crime data, three types of idea, type of crime, location of time and are fed into computer algorit algorithm using these areas are 500 by 500 square foot areas, maybe a city block and police are handed maps with these red boxes and they go patrol them. The logic is not algorithmic magic, its that some crime is contagious. People burglarize some house. Some car theft rings take place in various parking lots, its some environmentally vulnerable blt. Theyre the specific programs that add in the time of the day, the weather, to be able to forecast elevated risk in particular neighborhood and particular time. The result isnt particular of crime sz much where people can control. Secondarily theres subject based predictive policing. Chicagos Police Department has heat list, top 400 people who they think are more likely to be the victim of Violent Crime or perpetrator of Violent Crime. Those 400 people on the socalled heat list get a knock on the door. Yes, we know youre involved, go left, go to jail. And idea of targeting these people is now part of chicagos policing. In the minute i have left, i want to address how Constitutional Rights might be effected by these types of predictive policing. It protects all of us from unlawful, unreasonable searches and stop, frisks and searches, so the question is if you are in that box at 500 by 500 square foot box and Police Officer has been told to be on the look out for suspicious behavior and seize you, have your right change. If youre on the heat list and you have been designated one of the top 400 people most likely to be involved in Violent Crime, should your Fourth Amendment protections changed. The answers are hard. Its not easy, but it goes back to the difficulty of trying to post it, why suppose so valuable, its shedding light on the darkness of predictive policing and thank you all for being here and my time is up. Oush Bay Area Black lives matter chapter and do direct action training through a couple of different groups across the country. Im going to talk about three things when it comes to predictive policing, a little bit around the data that was talked a little bit earlier, the money the money and lack of accountability and whats happening on the ground, right, so the first one is the assumption that the data is mutual, which we know is not the case, right, we know we have historical studies that have proven over and over again that various communities might predict and if thats the case and thats the data thats going into these maps and going into the studies and its only going to continue to produce an Accurate Information that relies on racist human assumptions the other one is that theres an influx of money it was discussed earlier on one of the panels, around the sting ray one, in particular, the amount the amount of money thats coming in from the seizure of assets from Police Departments and also from the federal government over the last 13 years, its been about 23 million thats come in through the department of justice, to grant to local Police Departments, four different technological policing solutions and a lot of that has gone into predictive softwares that are unproven. At this point there are several departments that are saying actually dont work and because the money comes in through these grants, theyre not, then, given the same kind of accountability for the inflated Police Department budgets that were seeing, so we have Community Members that are fighting for Community Policing or other kind of policing structures and their cities, but millions and millions are being funneled for technologies that dont have the same level of accountability. And then the last one, really is around like whats happening on the ground, so we have all of these algorithms that are used to really predict crime and fight crime, what that means and whats not discussed is that its a continuation of the materialization of our Police Departments, so come from lapd spying spoke earlier around the use of drones and the militarization but these predictive policing practices and technologies were created for the military. They were created to be able to find hot spots and iraq and afghanistan and are now being used this our departments and Police Departments across the country because it can take public dollars into corporate into corporate into large corporations without the level of accountability and transparency that is required, generally for Police Departments. The other thing that i would say is that there are a lot of different groups that are kind of starting to work on this, right, so you have like baker center based out of oakland that is working to produce technology that really, then, is around Holding Police accountable. We had the technology earlier that brandon spoke around s. W. A. T. Thats around Police Accountability and police transparency, there are ways in which community are responding using technological mechanisms to be able to hold police accountable, but dont necessarily know as much around these the ways in which predictive policing and predictive sentencing and the big data is being used against us. So i would say, the last thing that ive seen in just in conversations that ive had with activists and organizers aacross the country in a lot of different communities is really, weve got to have a sense of it, like, we know that theres money in it, we know that because of the work of organizers across the country, police are theres been a massive spotlight on policing in this country. We know that and because of that, the answer has been how can we use technology to be able to solve that and how can we take the human nature out of it and, instead, using technology to really reproduce historical historical harms and using the technology to really reproduce structural racism. Thank you. And well conclude with christy who will tell us a little bit about how the government means in order to hand this will. Good afternoon, i want to thank Georgetown School of law for inviting me here to this important event, i think its most critical at this juncture within the United States. Crime fighting technology, intel lens led policing and more to inform critical prevention and Crime Prevention tactics. Although this term of art, the types of detective work to solve crimes is not. While there are enhanced technical capabilities, theres a distinct need to ensure that Civil Liberties predictions are adequately protected, i want to address and sum up some of the points that some of the prior speakers have made with respect to some of the funding and ill do that in just a minute. The term, predictive policing are Data Analytics raises fears, understandably so, that the police might engage in illegal practices that they may over step their bound and potentially use that information in an intelligence way that bridges Fourth Amendment as well as the First Amendment in other privacy laws. Simply put, the term conjures up images of minority support society where people are able to rest individual before a crime is committed. Making predictions is one half of prediction led policing and the other part is acting on those predictions. Its important to understand and under score the relationship between federal Law Enforcement. The majority of crimes are enforced at the state and local level. Within the schematic, it is important for the federal government to provide strong policy considerations that protect communities of color and to ensure that they are not negatively impacted or targeted on the bases of race. There is the need for clearly defined mission in order to understand what information should be collected is clear. In december of 2014, then attorney general holder updated the department of justices use of race policy. This policy bail builds upon and expands the framework of 2003 guidance and reaffirms the governments deep commitment to ensure the Law Enforcement agencies conduct activity in unbias manner. Bias practices as the federal government long recognized including what mr. Jim baker talked about, perpetuate and vitally important, biases practices are ineffective from Law Enforcement perspective. As former attorney general eric holder is stated, such practices are simply not good Law Enforcement. As Privacy Professionals and attorneys, we seek to distinguish information which determines what is and what is not protected under federal policy laws. In this day, privacy officials need to be engaged about what information can and to be shared with other agencies. The department of justice, these responsibilities are within various office within the department and we take a multi layered approach protecting privacy, Civil Liberties and civil rights. As mentioned, in 2014, the white house released a report titled big data seizing opportunities, preserving values and this was the result of president obamas 90day big day report. This report caught on the department of justice to examine and focus the domestic use of policing and specifically to focus on the more complex predictive technologies. For example, not merely mapping crimes have occurred in searching databases for specific information, as we know, the data is only as valuable and as good as it is collected, so if youre relying on older data and not raw data that is action nabl, youre not going to get the results that Law Enforcement needs to work in their Law Enforcement mission. So one of the things that was important within this report and within this tasks was that the department established guidelines for the use of state and local Law Enforcement because as we know, a lot of this is targeted at the level. Although the federal government, because federal and other issues cant specifically tell state and local Law Enforcement how to conduct the business, we cant establish policies and standard of conduct that can be used as a milestone or model for future policing activities, thank you. [ applause ] thank you, everyone. Id like to start off by just asking a very blunt question, and that is, predictive policing and sentencing racial profiling by another name. And id like to direct that question to janel first. Yeah, hello. So just the short answer of that, i would say yes. If we are studying the ways theres two parts, right, so the first one is the fact that it does rely on old data that is not neutral, right, like that is really the crux of it. The other one is that very similar to what sonja said we have these algorithms that are propry ti proprietary. I would say, between the two of those, it is, it is being able to take communities that are over policed and now finding new reasons and neutral reasons to police them again and police them more. Christy, did you have a response . I think one thing that we havent explicitly talked about is the need for transparency. And when president obama took office, one of the things that he mandated was open government and open data. And i think that one of the things that the government can do in a real concrete way is to require within certainly that are required within the federal government, but, basically, encourage the state and local Law Enforcements to have some sort of privacy policy, some sort of privacy Impact Assessment that lets the public know how the sims are developed, whether or not theyre proprietary and what are some of the risks in using that data to make sure theyre despaired impacts with respect to different communities being targeted or not targeted. Regarding sentencing policing. I think in terms of transparen transparency, the key issue that we need to focus on and what happens with discussions like this, is that we can begin demanding that. There isnt transparency right now. There are certain proprietary predictive policing technologies that dont share the information. And there are some that are recognizing that they need to and they will. But the government the Police Control the are not giving it up and i think because of that, its easy to label things as racial profiling, whe