Big Data Analytics in Law Enforcement Case Study: LASER Predictive Policing Program in Los Angeles

Submitted on 19/10/18 for the Technology Dynamics course. Grade: 8.5/10.

–

Big data analytics combines two concepts: big data and data analytics. Big data refers to an enormous amount of data, both structured and unstructured, which can produce valuable insights when combined with data analytics – a process of applying algorithms in examining data to find patterns, relationships, and information (Elgendy & Elragal, 2014). Big data analytics has various usage from predicting consumer’s shopping behavior, mitigating natural disasters, developing governmental programs, enforcing law, to preventing terrorism (Tene & Polonetsky, 2013). Big data analytics is a growing field of research, and with technological advancements and the ever-increasing size of data being produced every day, different tools and methods are still being developed by various innovation agents including private companies such as Microsoft, IBM, Tableau, and Palantir (SharesPost, Inc., 2017).

In order to increase efficiency and accuracy of big data analytics, different kinds of tools and methods have been developed, such as social media analytics, social network analysis (SNA), text mining, sentiment analysis, visual discovery and Advanced Data Visualization (ADV) (Elgendy & Elragal, 2014). Palantir, for instance, develops a platform that can integrate anydata sets from different types, visualize them in dynamic settings – helping human analysts analyze, find patterns, conclude, and create follow-up plans based on their findings (The Forward Operational Assessment (FOA) XVIII Team, 2012). The platform focuses on synergy between humans and computers, letting users ask relevant questions about the data using various analytical tools within the system – augmenting human intelligence instead of replacing it (Payne et al., 2008).

The benefits of big data analytics are not only being experienced by corporations (Davenport & Dyché, 2013), but also by the government (Morabito, 2015). However, it is argued that the benefits come at the cost of individuals’ privacy (Richards & King, 2013). The society has been increasingly aware of the importance of safeguarding data privacy, it is estimated that a total of 134 out of 195 countries will have data privacy laws that meet international standards by 2020, compared to 76 countries in 2011, and 109 countries in 2015 (Greenleaf, 2017). The rules and regulations regarding data privacy could hamper further development of big data analytics, making it the reverse salient of this technology. One of the function of reverse salient is that is pushes for continuous innovation (Dedehayir & Mäkinen, 2008), where in this context, the rules and regulations demand technology developers to ensure responsible usage of the data they gather and process.

The utilization of big data analytics in law enforcement has facilitated the shift of policing approach from reactive to preventive, an approach that is often associated with intelligence agencies (Završnik, 2013). Unlike the usage of big data analytics in corporations where data privacy can be safeguarded by anonymizing (or de-identifying) data to prevent tracing the original data owner (Duhigg, 2012), the analytics performed by law enforcement agencies rely on the attributes and personal information assigned on individuals, building a network of information on potential crime offenders (Brayne, 2017). This abstract will further discuss the implementation of big data analytics in law enforcement, and how it should be improved to avoid the risk of abuse and violation of privacy and civil liberty.

According to Bergek et al. (2008), there are three components of a Technological Innovation Systems; actors, networks, and institutions. The analysis in this abstract would focus on the geographical area of Los Angeles, the U.S. State of California, where big data analytics platform Palantir has been used to aid the development of a predictive policing program called LASER (Los Angeles’ Strategic Extraction and Restoration Program) since 2011 (LAPD & Uchida, 2014). The actors involved in this system include Palantir as technology developer, Los Angeles Police Department (LAPD) as technology adopter, California Senate as a governing body, and LA citizens as the customer of both LAPD and California State Senate.

The actors are all connected in the network through interactions, where Palantir develop various analytical tools in accordance to the need of LAPD, and the development of new regulations by governing body concerning the technology pressured by public protests (Macias, 2018). The institutional component includes national law on data privacy and SB-21, a senate bill that was amended in August 2017. The bill requires each law enforcement agency to attend regular public hearing to propose the usage of surveillance technology and the collected information, and publish reports concerning the use of the technology. This bill increases the transparency of LAPD programs, as well as to include civilians as an actor that could actively shape the development of the technology, and not only be affected by it.

According to Value Sensitive Design (VSD) approach, a technology is greatly influenced by human values (Friedman et al., 2002), and that is why it is important to first identify values of different stakeholders before and during the development of the technology. The citizens of Los Angeles value both public and personal security, as well as their personal privacy and liberty. As a technology developer, Palantir values its ability to accommodate customers’ needs while maintaining privacy & liberty. LAPD aims to maintain public security, reduce crime rate, and serve the general public. As a governing body, California State Senate’s values are creating law and regulations according to public needs.

However, LAPD has been maintaining public security at the cost of individual privacy; using external data that were not collected for the purposes of criminal justice such as electronic toll pass data, address and usage from utility bills, and even pizza delivery orders (Brayne, 2017). The system is also connected to other local, state, and federal data sources such as the Automated License Plate Recognition (ALPR) (Palantir Technologies Inc., 2014). This creates problem, because the connection to ALPR database can put people with no criminal justice contact into the law enforcement database – putting them under the radar (Brayne, 2017). The usage of ALPR and external data sources by the LAPD violates the two main principles on analyzing data for the purpose of law enforcement, as proposed by Cocx (2009): data on unsuspected individuals should only be analyzed in exceptional situations, and that it should only be analyzed for the purpose it was collected for.

Although Palantir has embed built-in preventive measure regarding privacy (Palantir Technologies Inc., 2012), it still does not guarantee the responsible usage of the platform. Other than that, LAPD has a slow response rate of public request for information (LAPD, 2018). Accordingly, LA citizens demand an audit of the predictive policing program (Macias, 2018), indicating a distrust of the system. Different steps have been taken to create a more responsible usage of big data analytics in law enforcement in Los Angeles, but it is still hard for both the police and the citizen to fully embed their conflicting values into the technology. The police need data on its citizens to monitor and detect potential crime offenders in order to ensure public security, but the citizens also need their privacy and personal freedom to be guaranteed, as both are the responsibility of the state (Broeders et al., 2017). It is apparent that a compromise needs to be made between the two stakeholders, with the presence of governing body to facilitate interaction and construct laws.

The usage of big data analytics platform in law enforcement has several negative implications such as the violation of privacy and civil liberties. Sensitive data are being collected without consent, and the analysis are being used to label civilians as a threat to the society’s security – despite the fact that they are also a part of the entity whose security should be guaranteed by the police. Furthermore, the data were analyzed by human analysts, who might have their own personal prejudices towards people from certain ethnicity and race, resulting in biased analysis (Tene & Polonetsky, 2013; James, 2017). The nature of data collection of the system creates a self-perpetuating cycle, where past interactions with police further increases the probability of future interactions with police, resulting in a narrowing target of police operations and discomfort to the targeted population.

To ensure unbiased analysis, there needs to be an evaluation of all data that are being used by LAPD and possibly exclude racial information. Other than that, the transparency of the system needs to be increased to gain trust from the public. One possible solution is to add a real-time auditing process performed by artificial intelligence to avoid human intervention. It should be developed in-house by the government instead of a private company, to prevent conflict of interests. To further increase the benefit of big data analytics for the society, it should not only be used to create crime-preventive program in a coercive manner such as LASER, but also crime-preventive program that focuses more on helping people finding decent employment so they won’t resort to crime. Big data analytics can also be used to develop rehabilitative measures with a goal to reinstate the crime offenders back into the society and prevent future offense. Instead of perceiving people as potential offenders, big data analytics could be a means for the government to manage its citizens in a more agreeable way.

References

• An act to add Chapter 15 (commencing with Section 54999.8) to Part 1 of Division 2 of Title 5 of the Government Code, relating to law enforcement agencies. (2017). SB-21.California.

• Bergek, A., Hekkert, M., & Jacobsson, S. (2008). Analyzing the Dynamics and Functionality of Technological Systems: A Manual. Research Policy , 37 (3), 407-429.

• Brayne, S. (2017). Big Data Surveillance: The Case of Policing. American Sociological Review , 82 (5), 977-1008.

• Broeders, D., Schrijvers, E., van der Sloot, B., van Brakel, R., de Hoog, J., & Ballin, E. H. (2017). Big Data and security policies: Towards a framework for regulating the phases of analytics and use of Big Data. Computer Law & Security Review 33 , 33, 309-323.

• Cocx, T. (2009). Algorithmic Tools for Data-Oriented Law Enforcement. NWO.

• Davenport, T. H., & Dyché, J. (2013, May). Big Data in Big Companies. Retrieved 2018, from https://docs.media.bitpipe.com/io_10x/io_102267/item_725049/Big-Datain-Big-Companies.pdf

• Dedehayir, O., Mäkinen, S.J. (2008). Dynamics of Reverse Salience as Technological Performance Gap: An Empirical Study of the Personal Computer Technology System. Journal of Technology Management & Innovation, 3(3), 55-66.

• Duhigg, C. (2012, February 16). How Companies Learn Your Secrets. Retrieved 2018, from The New York Times Magazine: https://www.nytimes.com/2012/02/19/magazine/shopping-habits.html

• Elgendy, N., & Elragal, A. (2014). Big Data Analytics: A Literature Review Paper. Lecture Notes in Computer Science (pp. 214-227). Springer International Publishing Switzerland.

• Friedman, B., Kahn Jr., P. H., & Borning, A. (2002). Value Sensitive Design: Theory and Methods. Universtiy of Washington Technical Report , 02-12.

• Greenleaf, G. (2017, January 30). Global Data Privacy Laws 2017: 120 National Data Privacy Laws, Including Indonesia and Turkey. 145 Privacy Laws & Business International Report, 10-13.

• James, L. (2017). The Stability of Implicit Racial Bias in Police Officers. Police Quarterly, 21 (1), 30-52.

• LAPD, I. T., & Uchida, C. (2014). Smart Policing: Institutionalizing Operation LASER in The Los Angeles Police Department. LAPD.

• Los Angeles Police Department. (2018, March 9). Stop LAPD Spying. Retrieved 2018, from Stop LAPD Spying: https://stoplapdspying.org/wpcontent/uploads/2018/05/Khan-Hamid-C17-0500027-Operation-LASER-1.pdf

• Macias, J. M. (2018, August 14). Retrieved 2018, from Courthouse News Services: https://www.courthousenews.com/lapd-officials-promise-review-of-techbased-policing-methods/

• Macias, J. M. (2018, May 8). Retrieved 2018, from Courthouse News Services: https://www.courthousenews.com/activists-call-for-an-end-to-laspredictive-policing-program/

• Morabito, V. (2015). Big Data and Analytics – Strategic and Organizational Impacts. Springer International Publishing Switzerland.

• Palantir Technologies Inc. (2012). A Core Commitment: Protecting Privacy and Civil Liberties. Retrieved 2018, from Palantir: https://www.palantir.com/_ptwp_live_ect0/wpcontent/uploads/2012/06/ProtectingPrivacy_CivilLiberties_2012.pdf

• Palantir Technologies Inc. (2014). Palantir Audit Logging. Retrieved 2018, from Palantir: https://www.palantir.com/wp-assets/media/capabilities-perspectives/Palantir-Audit-Logging.pdf

• Palantir. (2014, 3). Reponding to Crime in Real Time: Palantir at the Los Angeles Police Department. Retrieved 2018, from Palantir: https://www.palantir.com/wpassets/wp-content/uploads/2014/03/Impact-Study-LAPD.pdf

• Payne, J., Solomon, J., Sankar, R., & McGrew, B. (2008). Grand Challenge Award:Interactive Visual Analytics. Palantir: The Future of Analysis. IEEE Symposium on Visual Analytics Science and Technology (pp. 201-202). Columbus: IEEE.

• Richards, N. M., & King, J. H. (2013). Three Paradoxes of Big Data. Stanford LawReview Online 41 , 41-46.

• SharesPost, Inc. (2017). Company Report Palantir: Redefining Analytics, Augmenting Intelligence, & Unlocking Secrets. SharesPost Research LLC.

• Tene, O., & Polonetsky, J. (2013). Big Data for All: Privact and User Control in the Age of Analytics. Northwestern Journal of Technology and Intellectual Property, 11 (5), 240-273.

• The Forward Operational Assessment (FOA) XVIII Team. (2012). Forward Operational Assessment. Palantir: Operational Assessment Report

• Završnik, A. (2013). Blurring the Line between Law Enforcement and Intelligence: Sharpening the Gaze of Surveillance? Journal of Contemporary European Research, 9 (1), 181-202.