1932

Abstract

There are widespread concerns about the use of artificial intelligence in law enforcement. Predictive policing and risk assessment are salient examples. Worries include the accuracy of forecasts that guide both activities, the prospect of bias, and an apparent lack of operational transparency. Nearly breathless media coverage of artificial intelligence helps shape the narrative. In this review, we address these issues by first unpacking depictions of artificial intelligence. Its use in predictive policing to forecast crimes in time and space is largely an exercise in spatial statistics that in principle can make policing more effective and more surgical. Its use in criminal justice risk assessment to forecast who will commit crimes is largely an exercise in adaptive, nonparametric regression. It can in principle allow law enforcement agencies to better provide for public safety with the least restrictive means necessary, which can mean far less use of incarceration. None of this is mysterious. Nevertheless, concerns about accuracy, fairness, and transparency are real, and there are tradeoffs between them for which there can be no technical fix. You can't have it all. Solutions will be found through political and legislative processes achieving an acceptable balance between competing priorities.

Loading

Article metrics loading...

/content/journals/10.1146/annurev-criminol-051520-012342
2021-01-13
2024-04-24
Loading full text...

Full text loading...

/deliver/fulltext/criminol/4/1/annurev-criminol-051520-012342.html?itemId=/content/journals/10.1146/annurev-criminol-051520-012342&mimeType=html&fmt=ahah

Literature Cited

  1. Ackerman E. 2017. This autonomous swarm doesn't need GPS. IEEE Spectrum Dec. 27. https://spectrum.ieee.org/automaton/robotics/drones/this-autonomous-quadrotor-swarm-doesnt-need-gps
    [Google Scholar]
  2. Aldor-Noiman S, Brown LD, Fox EB, Stine RA 2016. Spatio-temporal low count processes with applications to violent crime events. Stat. Sin. 26:1587–610
    [Google Scholar]
  3. Angrist JD, Pischke J-S. 2009. Most Harmless Econometrics Princeton, NJ: Princeton Univ. Press
  4. Angwin A, Larson J, Mattui S, Kirchner L 2016. Machine bias. ProPublica May 23. https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing
    [Google Scholar]
  5. Anselin L, Cohen J, Cook D, Goor W, Tita G 2000. Spatial analyses of crime. Criminal Justice 2000: Measurement and Analysis of Crime and Justice R Kaminsk, N La Vigne 213–62 Washington, DC: Natl. Inst. Justice
    [Google Scholar]
  6. Arute F, Arya K, Babbuch R, Bacon D, Bardin JC et al. 2019. Quantum supremacy using a programmable superconducting processor. Nature 575:505–10
    [Google Scholar]
  7. Bennett Moses L, Chan J 2018. Algorithmic prediction in policing: assumptions, evaluation, and accountability. Polic. Soc. 28:7806–22
    [Google Scholar]
  8. Berk RA. 2017. An impact assessment of machine learning risk forecasts on parole board decisions and recidivism. J. Exp. Criminol. 13:193–216
    [Google Scholar]
  9. Berk RA. 2018. Machine Learning Forecasts of Risk in Criminal Justice Settings Cham, Switz.: Springer
  10. Berk RA. 2020a. Almost politically acceptable criminal justice risk assessment. Criminol. Public Policy. In press
    [Google Scholar]
  11. Berk RA. 2020b. Statistical Learning from a Regression Perspective Cham, Switz.: Springer, 3rd ed..
  12. Berk RA, Bleich J. 2013. Statistical procedures for forecasting criminal behavior: a comparative assessment. J. Criminol. Public Policy 12:3515–44
    [Google Scholar]
  13. Berk RA, Brown L, Buja A, George E, Pitkin E et al. 2014. Misspecified mean function regression: making good use of regression models that are wrong. Sociol. Methods Res. 43:433–51
    [Google Scholar]
  14. Berk RA, Buja A, Brown L, George E, Kuchibhotla AK et al. 2019. Assumption lean regression. Am. Stat. https://doi.org/10.1080/00031305.2019.1592781
    [Crossref] [Google Scholar]
  15. Berk RA, Heidari H, Jabbari S, Kearns M, Roth A 2018. Fairness in criminal justice risk assessment: the state of the art. Sociol. Methods Res. https://doi.org/10.1177/0049124118782533
    [Crossref] [Google Scholar]
  16. Berk RA, Sorenson SB. 2016. Forecasting domestic violence: a machine learning approach to help inform arraignment decisions. J. Empir. Legal Stud. 13:195–115
    [Google Scholar]
  17. Berk RA, Sorenson SB. 2020. An algorithmic approach to forecasting rare violent events: an illustration based on intimate partner violence perpetration. Criminol. Public Policy 19:1213–33
    [Google Scholar]
  18. Berk RA, Sorenson SB, He Y 2005. Developing a practical forecasting screener for domestic violence. Eval. Rev. 29:4358–82
    [Google Scholar]
  19. Bowers KJ, Johnson SD, Pease K 2004. Prospective hot-spotting: the future of crime mapping. Br. J. Criminol. 44:541–658
    [Google Scholar]
  20. Box GEP, Jenkins GW, Reinsel GC, Ljung GM 2016. Time Series Analysis: Forecasting and Control Hoboken, NJ: Wiley
    [Google Scholar]
  21. Braga AA, Papachristos AV, Hureau DH 2014. The effects of hot spots policing on crime: an updated systematic review and meta-analysis. Justice Q 31:4633–63
    [Google Scholar]
  22. Breiman L. 2001a. Random forests. Mach. Learn. 45:5–32
    [Google Scholar]
  23. Breiman L. 2001b. Statistical modeling: two cultures (with discussion). Stat. Sci. 16:199–231
    [Google Scholar]
  24. Breiman L, Friedman JH, Olshen RA, Stone CJ 1984. Classification and Regression Trees Monterey, CA: Wadsworth Press
  25. Burgess EM. 1928. Factors determining success or failure on parole. The Working of the Indeterminate Sentence Law and the Parole System in Illinois AA Bruce, AJ Harno, EW Burgess, EW Landesco 205–49 Springfield, IL: State Board Parole
    [Google Scholar]
  26. Cameron AC, Trivedi PK. 2005. Macroeconometrics: Methods and Applications New York: Cambridge Univ. Press
    [Google Scholar]
  27. Caplan JM. 2011. Mapping the spatial influence of crime correlates: a comparison of operational schemes and implications for crime analysis and criminal justice practice. J. Policy Dev. Res. 13:357–83
    [Google Scholar]
  28. Caplan JM, Kennedy LW, Miller J 2010. Risk terrain modeling: brokering criminological theory and GIS methods for crime forecasting. Justice Q 28:2360–81
    [Google Scholar]
  29. Caplan JM, Kennedy LW. 2016. Risk Terrain Modeling: Crime Prediction and Risk Reduction Berkeley: Univ. Calif. Press
    [Google Scholar]
  30. Cellen-Jones R. 2014. Stephen Hawking warns artificial intelligence could end mankind. BBC News: Technology Dec. 2. https://www.bbc.com/news/technology-30290540
    [Google Scholar]
  31. Chinoy S. 2019. The racist history behind facial recognition. New York Times July 10. https://www.nytimes.com/2019/07/10/opinion/facial-recognition-race.html
    [Google Scholar]
  32. Chouldechova A. 2017. Fair prediction with disparate impact: a study of bias in recidivism prediction instruments. arXiv:1610.075254v1 [stat.AP]
  33. Coglianese C, Lehr D. 2017. Regulating by robot: administrative decision making in the machine-learning era. Georget. Law J. 105:1147–223
    [Google Scholar]
  34. Corbett-Davies S, Goel S. 2018. The measure and mismeasure of fairness: a critical review of fair machine learning Paper presented at the 35th International Conference on Machine Learning Stockholm:
    [Google Scholar]
  35. Cressie NAC. 1993. Statistics for Spatial Data Hoboken, NJ: Wiley
  36. Dinis-Oliveira RJ. 2017. Metabolic profile of flunitrazepam: clinical and forensic toxicological aspects. Drug Metab. Lett. 11:114–20
    [Google Scholar]
  37. Domingos P. 2015. The Master Algorithm: How the Quest for the Ultimate Learning Machine Will Remake Our World New York: Basic Books
    [Google Scholar]
  38. D'Orsogna MR, Perc M. 2015. Statistical physics of crime: a review. Phys. Life Rev. 12:1–21
    [Google Scholar]
  39. Dwork C, Roth A. 2014. The algorithmic foundations of differential privacy. Found. Trends Theor. Comput. Sci. 9:3–4211–407
    [Google Scholar]
  40. Ferguson AG. 2017. The Rise of Big-Data Policing: Surveillance, Race, and the Future of Law Enforcement New York: NYU Press
    [Google Scholar]
  41. Flaxman S, Chirico M, Pereira P, Loeffler C 2019. Scalable high-resolution forecasting of sparse spatiotemporal events with kernel methods: a winning solution to the NIJ “Real-Time Crime Forecasting Solution. .” Ann. Appl. Stat. 13:42564–85
    [Google Scholar]
  42. Freedman DA. 2009. Statistical Models: Theories and Practice Cambridge, UK: Cambridge Univ. Press, 2nd ed..
    [Google Scholar]
  43. Friedman JS. 2002. Stochastic gradient boosting. Comput. Stat. Data Anal. 38:367–78
    [Google Scholar]
  44. Goel S, Shroff R, Skeem JL, Slobogin C 2019. The accuracy, equity, and jurisprudence of criminal risk assessment. Risk-Resil. Res. In press. http://risk-resilience.berkeley.edu/sites/default/files/journal-articles/files/manuscript_the_accuracy_equity_and_jurisprudence_of_criminal_risk_assessment_1.2.19_.pdf
    [Google Scholar]
  45. Goodfellow I, Bengio Y, Courville A 2016. Deep Learning Cambridge, MA: MIT Press
  46. Grahm O, Dowe D. 2019. The Turing test. The Stanford Encyclopedia of Philosophy EZ Zalta Stanford, CA: CSLI
    [Google Scholar]
  47. Groff ER, La Vigne NG 2004. Forecasting the future of predictive crime mapping. Crime Prev. Stud. 13:29–57
    [Google Scholar]
  48. Hamilton M. 2016. Risk-needs assessment: constitutional and ethical challenges. Am. Crim. Law Rev. 52:2231–92
    [Google Scholar]
  49. Hastie T, Tibshirani R, Friedman J 2009. The Elements of Statistical Learning: Data Mining, Inference, and Prediction New York: Springer, 2nd ed..
    [Google Scholar]
  50. Helmstetter A, Ouillon G, Sornette D 2003. Are aftershocks for large California earthquakes diffusing. J. Geophys. Res. 108:B102483
    [Google Scholar]
  51. Hill K. 2020. Activate this ‘bracelet of silence,’ and Alexa can't eavesdrop. New York Times Febr. 16, Sect. BU 3
    [Google Scholar]
  52. Hollywood JS, McKay KN, Woods D, Agneil D 2019. Real-Time Crime Centers in Chicago Santa Monica, CA: RAND
  53. Hu Y, Wang F, Guin C, Zhu H 2018. A spatial-temporal kernel density estimation framework for predictive crime hotspot mapping and evaluation. Appl. Geogr. 9:89–97
    [Google Scholar]
  54. Huq AZ. 2019. Racial equality in algorithmic criminal justice. Duke Law J 68:61043–134
    [Google Scholar]
  55. Johnson SD, Birks DJ, McLaughlin L, Bowers KJ, Pease K 2007. Prospective crime mapping in operational context Home Off. Rep. 19/07, Res. Dev. Stat. Dir London:
  56. Kaufmann M, Egbert S, Leese M 2019. Predictive policing and the politics of patterns. Br. J. Crimiol. 59:674–92
    [Google Scholar]
  57. Kearns M, Neel S, Roth A, Wu Z 2018a. An empirical study of rich subgroup fairness for machine learning. arXiv:1808.08166v1 [cs.LG]
  58. Kearns M, Neel S, Roth A, Wu Z 2018b. Preventing fairness gerrymandering: auditing and learning subgroup fairness. arXiv 1711.05144v4 [cs.LG]
  59. Kearns M, Roth A. 2019. The Ethical Algorithm: The Science of Socially Aware Algorithm Design Oxford, UK: Oxford Univ. Press
    [Google Scholar]
  60. Kessel JM. 2019. Killer robots aren't regulated. Yet. New York Times Dec. 13. https://www.nytimes.com/2019/12/13/technology/autonomous-weapons-video.html
    [Google Scholar]
  61. Kleinberg J, Mullainathan S, Raghavan M 2017. Inherent tradeoffs in the fair determination of risk scores Paper presented at the 8th Innovations in Theoretical Computer Science Conference Berkeley:
  62. Krohn J. 2019. Deep Learning Illustrated: A Visual, Interactive Guide to Artificial Intelligence Boston: Addison-Wesley
    [Google Scholar]
  63. Kroll JA, Huey J, Barocas S, Felten EW 2017. Accountable algorithms. Univ. Pa. Law Rev. 165:3633–705
    [Google Scholar]
  64. Lei J, G'Sell M, Rinaldo A, Tibshirani RJ, Wasserman L 2018. Distribution-free predictive inference for regression. J. Am. Stat. Assoc. 113:5231094–111
    [Google Scholar]
  65. Liesenfeld R, Richard J-F, Vogler J 2017. Likelihood-based inference and prediction in spatio-temporal panel count models of urban crimes. J. Appl. Econom. 32:500–620
    [Google Scholar]
  66. Lyngstad T, Skardhammar T, Berk R 2017. Predicting future crime at birth Paper presented at the 73rd Annual Meeting of the American Society for Criminology Philadelphia:
  67. Malkevitch J. 2007. Taxi! American Mathematical Society October. https://www.ams.org/publicoutreach/feature-column/fcarc-taxi
    [Google Scholar]
  68. Mayson SG. 2019. Bias in, bias out. Yale Law J 128:2218–300
    [Google Scholar]
  69. Metz C. 2019. Google claims a quantum breakthrough that could change computing. New York Times Oct. 23. https://www.nytimes.com/2019/10/23/technology/quantum-computing-google.html
    [Google Scholar]
  70. Mitchell M. 2019a. Artificial Intelligence: A Guide for Thinking Humans New York: Farrar, Straus and Giroux
    [Google Scholar]
  71. Mitchell M. 2019b. We shouldn't be scared of “superintelligent A.I.”. New York Times Oct. 31. https://www.nytimes.com/2019/10/31/opinion/superintelligent-artificial-intelligence.html
    [Google Scholar]
  72. Mohler G. 2014. Marked point process hotspot maps for homicide and gun crime prediction in Chicago. Int. J. Forecast. 30:3491–97
    [Google Scholar]
  73. Mohler GO, Short MB, Brantingham PJ, Schoenberg FP, Tita GE 2011. Self-exciting point process modeling of crime. J. Am. Stat. Assoc. 493:100–8
    [Google Scholar]
  74. Mohler GO, Short MB, Malinowski S, Johnson M, Tita GE et al. 2015. Randomized controlled field trials of predictive policing. J. Am. Stat. Assoc. 110:5121399–411
    [Google Scholar]
  75. Mullainathan S. 2019. Biased algorithms are easier to fix than biased people. New York Times Dec. 6. https://www.nytimes.com/2019/12/06/business/algorithm-bias-fix.html
    [Google Scholar]
  76. Munoz D, Bouchereau F, Vargas C, Enriquez R 2009. Heuristic approaches to the position location problem. Position Location Techniques and Applications103–52 Cambridge, MA: Academic
    [Google Scholar]
  77. Ohlin LE, Duncan OD. 1949. The efficiency of prediction in criminology. Am. J. Sociol. 54:441–52
    [Google Scholar]
  78. Pande V. 2018. Artificial intelligence's ‘black box’ is nothing to fear. New York Times Jan. 25. https://www.nytimes.com/2018/01/25/opinion/artificial-intelligence-black-box.html
    [Google Scholar]
  79. Pascanu R, Gulcehere C, Cho K, Begio Y 2014. How to construct deep recurrent neural networks. arXiv:13126026v5 [cs.NE]
  80. Perry WL, McInnis B, Price CC, Smith SC, Hollywood JS 2013. Predictive Policing: The Role of Crime Forecasting in Law Enforcement and Operations Santa Monica, CA: RAND
    [Google Scholar]
  81. Pindyck RS, Rubinfeld DL. 1981. Econometric Models and Economic Forecasts New York: McGraw Hill, 2nd ed..
  82. Poole D, Mackworth A, Goebel R 1998. Computational Intelligence: A Logic Approach Oxford, UK: Oxford Univ. Press
    [Google Scholar]
  83. Rachlinski JJ, Johnson SL, Wistrich AJ, Guthrie C 2009. Does unconscious racial bias affect trial judges. Notre Dame Law Rev 84:31195–252
    [Google Scholar]
  84. Ratcliffe JH. 2014. The hotspot matrix: a framework for spatio-temporal targeting of crime reduction. Police Pract. Res. 5:15–23
    [Google Scholar]
  85. Ratcliffe JH, McCullagh MJ. 2001. Chasing ghosts? Police perception of high crime areas. Br. J. Criminol. 41:330–41
    [Google Scholar]
  86. Reinhart A, Greenhouse J. 2018. Self-exciting point processes with spatial covariates: modelling the dynamics of crime. J. R. Stat. Soc. C 67:51305–29
    [Google Scholar]
  87. Reiss AJ. 1951. The accuracy, efficiency, and validity of a prediction instrument. Am. J. Sociol. 17:268–74
    [Google Scholar]
  88. Rey SJ, Mack EA, Koschinsky J 2012. Exploratory space-time analysis of burglary patterns. J. Quant. Criminol. 28:509–31
    [Google Scholar]
  89. Rosser G, Cheng T. 2016. Improving the robustness and accuracy of crime prediction with a self-exciting point process through isotropic triggering. Appl. Spat. Anal. 12:5–25
    [Google Scholar]
  90. Rosser G, Davies T, Bowers K, Johnson SD, Cheng T 2017. Predictive crime mapping: arbitrary grids or street networks?. J. Quant. Criminol. 33:569–94
    [Google Scholar]
  91. Scharre P. 2018. Army of None: Autonomous Weapons and the Future of War New York: Norton
    [Google Scholar]
  92. Smith CS. 2019. Dealing with bias in artificial intelligence: three women with extensive experience in A.I. spoke on the topic and how to confront it. New York Times Novemb. 19. https://www.nytimes.com/2019/11/19/technology/artificial-intelligence-bias.html
    [Google Scholar]
  93. Spielkamp M. 2017. Inspecting algorithms for bias. MIT Technology Review Jan. 12. https://www.technologyreview.com/s/607955/inspecting-algorithms-for-bias
    [Google Scholar]
  94. Starr S. 2014. Sentencing by the numbers. New York Times Aug. 10. https://www.nytimes.com/2005/01/02/magazine/sentencing-by-the-numbers.html
    [Google Scholar]
  95. Szkola J, Piza EL, Drawve G 2019. Risk terrain modeling: seasonality and predictive validity. Justice Q http://doi.org/10.1080/07418825.2019.1630472
    [Crossref] [Google Scholar]
  96. Tarling R, Perry JA. 1981. Statistical methods in criminological prediction. Prediction in Criminology DP Farrington, R Tarling 210–30 Albany, NY: SUNY Press
    [Google Scholar]
  97. Tibshirani R. 1996. Regression shrinkage and selection via the lasso. J. R. Stat. Soc. Ser. B 58:1267–88
    [Google Scholar]
  98. Turing A. 1950. Computing machinery and intelligence. Mind 59:237433–60
    [Google Scholar]
  99. Wang B, Yin P, Bertozzi AL, Brantingham PJ, Osher SJ, Xin J 2017a. Deep learning for real-time crime forecasting and its ternarization. arXiv:1711.08833v1 [cs.LG]
  100. Wang B, Zhang D, Zhang D, Brantingham PJ, Bettozzi AL 2017b. Deep learning for real time crime forecasting. arXiv:1707.03340v1 [math.NA]
  101. Weisburd D, Groff ER, Yang S-M 2012. The Criminology of Place Oxford, UK: Oxford Univ. Press
  102. Weisburd D, Mastrofski SD, Greenspan R, Willis JJ 2004. The growth of Compstat in American policing Rep., Police Found Washington, DC:
  103. Zeng J, Ustun B, Rudin C 2017. Interpretable classification models for recidivism prediction. J. R. Stat. Soc. Ser. A 180:3689–722
    [Google Scholar]
  104. Zhang Y, Cheng T. 2020. Graph deep learning model for network-based predictive hotspot mapping of sparse spatio-temporal events. Comput. Environ. Urban Syst. 79:101403
    [Google Scholar]
  105. Zhuang Y, Almeida M, Morabito M, Ding W 2017. Crime hot spot forecasting: a recurrent model with spatial and temporal information Paper presented at the 8th IEEE International Conference on Big Knowledge Hefei, China:
    [Google Scholar]
  106. Zucchini W, MacDonald IL, Langrock R 2016. Hidden Markov Models for Time Series: An Introduction Using R Boca Raton, FL: CRC Press, 2nd ed..
    [Google Scholar]
/content/journals/10.1146/annurev-criminol-051520-012342
Loading
/content/journals/10.1146/annurev-criminol-051520-012342
Loading

Data & Media loading...

  • Article Type: Review Article
This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error