Abstract
Fighting crime has historically been a field that drives technological innovation, and it can serve as an example of different governance styles in societies. Predictive policing is one of the recent innovations that covers technical trends such as machine learning, preventive crime fighting strategies, and actual policing in cities. However, it seems that a combination of exaggerated hopes produced by technology evangelists, media hype, and ignorance of the actual problems of the technology may have (over-)boosted sales of software that supports policing by predicting offenders and crime areas. In this paper we analyse currently used predictive policing software packages with respect to common problems of data mining, and describe challenges that arise in the context of their socio-technical application.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Zurich (Switzerland) and Munich (Germany) are using the Software Precobs to predict burglaries. The software is developed by IFMPT http://www.ifmpt.com/.
The Transcrime Research Center tried to predict burglaries in Rome, Milano and Bari for the year 2014. Report available online http://www.academia.edu/download/39022476/Transcrime_Research_in_Brief_Prevedere_i_furti_in_abitazione.pdf.
In France, the technology was evaluated in 2015 http://www.20minutes.fr/societe/1612375-20150521-viols-agressions-cambriolages-nouvel-algorithme-gendarmes-predire-crime.
See http://predpol.com/.
“Among the factors are the extent of a person’s rap sheet, his or her parole or warrant status, any weapons or drug arrests, his or her acquaintances and their arrest histories — and whether any of those associates have been shot in the past” (Gorner 2013).
This example has been heavily oversimplified for the purposes of a non-technical introduction. Any realistic classifier learning would take better account of noise, etc. The use of training and test data is also somewhat more involved in practice. Accessible (even if technical) introductions can be found in Witten et al. (2011). Their teaching materials are available at http://www.cs.waikato.ac.nz/ml/weka/book.html, see Chapter 5 on evaluation.
After 3 years a city with around 70,000 discontinued a contract with PredPol with the police chief commenting that “PredPol system may have greater benefit to law enforcement organisations policing much larger geographical jurisdictions where greater variables in crime patterns may exist”, while in Milpitas “existing internal processes of tracking crime and identifying potential areas of exposure were often more accurate than results received from PredPol”. (Ian Bauer, The Mercury News, 14.07.2016, accessible online at http://www.govtech.com/public-safety/Milpitas-Calif-Police-Department-Nixes-Predictive-Policing-Contract.html).
The evidence is cynically supported by the fact Chicago saw huge surge in number of homicides in 2016. http://crime.chicagotribune.com/chicago/homicides.
References
Anderson C (2008) The end of theory: the data deluge makes the scientific method obsolete. Wired Mag 16(7):16
Azavea (2015) HunchLab under the hood. https://cdn.azavea.com/pdfs/hunchlab/HunchLab-Under-the-Hood.pdf. Accessed 15 May 2017
Barocas S, Selbst AD (2016) Big Data’s disparate impact. 104 California Law Review 671. http://ssrn.com/abstract=2477899
Berendt B (2012) More than modelling and hiding: towards a comprehensive view of web mining and privacy. Data Min Knowl Discov 24(3):697–737
Biermann K (2015) Predictive policing: noch hat niemand bewiesen, dass data mining der Polizei hilft. Die Zeit, March 29, sec. Digital. http://www.zeit.de/digital/datenschutz/2015-03/predictive-policing-software-polizei-precobs/komplettansicht
Blumer A, Ehrenfeucht A, Haussler D, Warmuth MK (1990) Occam’s Razor. In: Readings in machine learning. pp 201–204
Caplan JM, Kennedy LW (2011) Risk terrain modeling compendium: for crime analysis. Rutgers Center on Public Security, New Jersey
Chainey S, Ratcliffe J (2013) GIS and crime mapping. Wiley, Hoboken
Citron DK (2007) Technological due process. Wash Univ Law Rev 85:1249–1313
Coudert F (2015) ‘Precrime police’ is not for 2054, it’s for now: how to regulate ‘data intensive policing’? In: Amsterdam privacy conference, Amsterdam, 23–26 October 2015
Drawve G (2014) A metric comparison of predictive hot spot techniques and RTM. Justice Q 33:1–29. doi:10.1080/07418825.2014.904393
ECJ/Court of Justice of the European Union (2016). Judgment of the Court (Grand Chamber) of 21 December 2016. Tele2 Sverige AB v Post- och telestyrelsen and Secretary of State for the Home Department v Tom Watson and Others. http://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:62015CJ0203
Eterno JA, Verma A, Silverman EB (2016) Police manipulations of crime reporting: insiders’ revelations. Justice Q 33(5):811–835. doi:10.1080/07418825.2014.980838
Ferguson AG (2015) Big Data and predictive reasonable suspicion. Univ Pennsylvania Law Rev 1632:327–410
Gorner J (2013) Chicago police use heat list as strategy to prevent violence. Chic Tribune 21:2013. http://articles.chicagotribune.com/2013-08-21/news/ct-met-heat-list-20130821_1_chicago-police-commander-andrew-papachristos-heat-list
Hunt P, Saunders J, Hollywood JS (2014). Evaluation of the shreveport predictive policing experiment. Research Report RR-531-NIJ. RAND Corp. http://www.rand.org/pubs/research_reports/RR531.html
Johnson SD (2008) Repeat burglary victimisation: a tale of two theories. J Exp Criminol 4(3):215–240. doi:10.1007/s11292-008-9055-3
Jouvenal J (2016) The new way police are surveilling you: calculating your threat ‘score’. The Washington Post, January 10. https://www.washingtonpost.com/local/public-safety/the-new-way-police-are-surveilling-you-calculating-your-threat-score/2016/01/10/e42bccac-8e15-11e5-baf4-bdf37355da0c_story.html
Lipton ZC (2016) The mythos of model interpretability. In: ICML 2016 workshop on human interpretability in machine learning (WHI 2016). http://zacklipton.com/media/papers/mythos_model_interpretability_lipton2016.pdf
Lum K, Isaac W (2016) To predict and serve? Significance 13(5):14–19
Mayer-Schönberger V, Cukier K (2013) Big Data: a revolution that will transform how we live, work and think. John Murray, London
McCarthy GF (2013) Custom notifications in Chicago—pilot program. 13–080 TRH. https://web.archive.org/web/20160414031801/, http://directives.chicagopolice.org/directives-mobile/data/a7a57bf0-13fa59ed-26113-fa63-2e1d9a10bb60b9ae.html
Meaney R (2004) Commuters and marauders: an examination of the spatial behaviour of serial criminals. J Investig Psychol Offen Profil 1(2):121–137. doi:10.1002/jip.12
Mohler G, Short M (2012) Geographic profiling from kinetic models of criminal behavior. SIAM J Appl Math 72(1):163–180. doi:10.1137/100794080
Mohler GO, Short MB, Brantingham PJ, Schoenberg FP, Tita GE (2012) Self-exciting point process modeling of crime. J Am Stat Assoc. doi:10.1198/jasa.2011.ap09546
Mohler G et al (2013) Modeling and estimation of multi-source clustering in crime and security data. Ann Appl Stat 7(3):1525–1539
Mohler GO, Short MB, Malinowski S, Johnson M, Tita GE, Bertozzi AL, Brantingham PJ (2015) Randomized controlled field trials of predictive policing. J Am Stat Assoc. doi:10.1080/01621459.2015.1077710
Morozov E (2013) To save everything, click here: technology, solutionism, and the urge to fix problems that don’t exist. Allen Lane, London
Papachristos AV (2009) Murder by structure: dominance relations and the social structure of gang homicide1. Am J Sociol 115(1):74–128. doi:10.1086/597791
Perry WL, McInnis B, Price CC, Smith S, Hollywood JS (2013) Predictive policing—the role of crime forecasting in law enforcement operations. Research Report RR-233-NIJ. RAND Corp. http://www.rand.org/pubs/research_reports/RR233.html
Saunders J, Hunt P, Hollywood JS (2016) Predictions put into practice: a quasi-experimental evaluation of Chicago’s predictive policing pilot. J Exp Criminol. doi:10.1007/s11292-016-9272-0
Short MB, D’Orsogna MR, Brantingham PJ, Tita GE (2009) Measuring and modeling repeat and near-repeat burglary effects. J Quant Criminol 25(3):325–339. doi:10.1007/s10940-009-9068-8
Short Martin B, Jeffrey Brantingham P, Bertozzi Andrea L, Tita George E (2010) Dissipation and displacement of hotspots in reaction-diffusion models of crime. Proc Natl Acad Sci 107(9):3961–3965. doi:10.1073/pnas.0910921107
Solove D (2011) Nothing to hide. The false trade-off between privacy and security. Yale University Press, Yale
Solove DJ (2006) A taxonomy of privacy. Univ Pa Law Rev 154(3): 477. GWU Law School Public Law Research Paper No. 129. https://ssrn.com/abstract=667622
Stroud M (2014) The minority report: Chicago’s new police computer predicts crimes, but is it racist? The Verge, 19 Feb 2014. http://www.theverge.com/2014/2/19/5419854/the-minority-report-this-computer-predicts-crime-but-is-it-racist
Townsley M, Homel Ross, Chaseling Janet (2003) infectious burglaries. A test of the near repeat hypothesis. Br J Criminol 43(3):615–633. doi:10.1093/bjc/43.3.615
Trotta M (2010) Serial offenders’ spatial behaviour: revisiting the marauder/commuter dichotomy. In: Presented at the 10th conference of the european society of criminology, Liège. Retrieved from http://orbi.ulg.ac.be/handle/2268/73623
Van Alsenoy B, Kuczerawy A, Ausloos J (2013) Search engines after ‘Google Spain’: Internet@Liberty or Privacy@Peril? TPRC 41. In: The 41st research conference on communication, information and internet policy. http://ssrn.com/abstract=2321494
Witten IH, Frank E, Hall MA (2011) Data mining. Practical machine learning tools and techniques, 3rd edn. Morgan Kaufmann, Burlington
Yang M, Coid J (2010) The efficacy of violence prediction: a meta-analytic comparison of nine risk assessment tools. Psychol Bull 136(5):740–767. doi:10.1037/a0020473
Zeng J, Ustun B, Rudin C (2016) Interpretable classification models for recidivism prediction. In: Presentation at FATML 2016. New York, 18 November 2017. https://arxiv.org/abs/1503.07810
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Degeling, M., Berendt, B. What is wrong about Robocops as consultants? A technology-centric critique of predictive policing. AI & Soc 33, 347–356 (2018). https://doi.org/10.1007/s00146-017-0730-7
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00146-017-0730-7