Abstract
While seeping into every aspect of our lives, intelligent systems carry serious concerns, many of which stem from the need to secure control over machines ensuring they turn into an enhancement and not into a threat to human societies. Issues are not lacking. Whether one considers recruiting platforms perpetuating historical patterns of discrimination or AI-driven systems unfairly evaluating the creditworthiness of natural persons, it clearly appears vital to firmly place human beings with their rights and needs at the center of intelligent technologies. The paper tackles this kind of issue by focusing on the use of artificial intelligence in criminal justice, where predictive analytics and AI-driven decision systems have proven capable not only of enhancing the fight against crime but also of bringing the risk of opacity, aberrations, and injustices. We present a human–machine collaboration strategy providing judges with the advantages of AI and computational heuristics while offering control and understanding of the role played by machines. Drawing from a research that led to the development of an experimental platform supporting judges and public prosecutors dealing with organized crime, the strategy revolves around two components: (i) an online learning model designed to support judges in the evaluation of criminal dangerousness of individuals and groups capable of learning from users’ feedback; (ii) a human–computer interaction component exploiting visual metaphors to ease judges’ interaction with data, AI and other heuristics. The main contribution of the work is a novel and viable declension of the Human-centered AI paradigm in justice administration.
Similar content being viewed by others
Data availability
Enquiries about data availability should be directed to the authors.
Notes
2012/C 326/02.
Evidence and facts emerging before the judge frequently call for “contextualized” assessment. For instance, the number or the direction of phone calls between individuals under surveillance is not unequivocally meaningful by itself. It may significantly vary depending on the organization considered, the region and the culture—in a way that only a human being can evaluate.
It is a statistical measure for assessing the reliability of agreement between a fixed number of raters when assigning categorical ratings to a number of items or classifying items; Fleiss’ kappa can range from \(-1\) to \(+1\), with \(-1\) indicating that there was no observed agreement, 0 indicating that agreement was no better than chance, and \(+1\) that indicates perfect agreement.
The choice can be made among individual-phonecall graph, individual-environmental tapping graph, individual crime graph, and combination thereof, as well as individual network.
With regards to the semantic meaning displayed, we have used, for example, influence for the betweenness centrality (\(f_8\) and \(F_{16}\)), popularity for (), gregariousness for ().
We remark that most of the systems PPs use in Italy, like the so called “Consolle del Magistrato” are mainly management tools with tabular visualization, that do not offer enhanced forms of interaction with data nor visual metaphors.
References
Aberdeen D, Pacovsky O, Slater A (2010) The learning behind gmail priority inbox
Allhutter D, Cech F, Fischer F, Grill G, Mager A (2020) Algorithmic profiling of job seekers in Austria: how austerity politics are made effective. Front Big Data. Vol. 5
Alves LG, Ribeiro HV, Rodrigues FA (2018) Crime prediction through urban metrics and statistical learning. Phys A Stat Mech Appl 505:435–443
Ashley KD (2017) Artificial intelligence and legal analytics: new tools for law practice in the digital age. Cambridge University Press, Cambridge
Bogomolov A, Lepri B, Staiano J, Oliver N, Pianesi F, Pentland A (2014) Once upon a crime: towards crime prediction from demographics and mobile data. In: Proceedings of the 16th international conference on multimodal interaction (pp. 427–434)
Carneiro D, Novais P, Andrade F, Zeleznikow J, Neves J (2014) Online dispute resolution: an artificial intelligence perspective. Artif Intell Rev 41(2):211–240
Carter S, Nielsen M (2017) Using artificial intelligence to augment human intelligence. Distill 2(12):e9
Choi JH, Hickman KE, Monahan A, Schwarcz D (2023) Chatgpt goes to law school. Available at SSRN
Cioffi-Revilla C (2014) Introduction to computational social science. Springer, London
Cortes C, Vapnik V (1995) Supportvector networks. Mach Learn 20(3):273–297
Delahoz-Dominguez EJ, Fontalvo-Herrera TJ, Mendoza-Mendoza AA (2020) Definición de perfiles geográficos de hurto de automóviles. caso aplicado en cartagena. Justicia 25(37):99–108
De Prisco R, Guarino A, Lettieri N, Malandrino D, Zaccagnino R (2021) Providing music service in ambient intelligence: experiments with gym users. Expert Syst Appl 177:114951
Esquivel N, Nicolis O, Peralta B, Mateu J (2020) Spatio-temporal prediction of baltimore crime events using clstm neural networks. IEEE Access 8:209101–209112
Floridi L, Chiriatti M (2020) Gpt-3: its nature, scope, limits, and consequences. Mind Mach 30(4):681–694
Gerritsen C, Elffers H (2020) Agent-based modelling for criminological theory testing and development. Routledge, Milton Park
Grabmair M, Ashley KD, Chen R, Sureshkumar P, Wang C, Nyberg E, Walker VR (2015) Introducing luima: an experiment in legal conceptual retrieval of vaccine injury decisions using a uima type system and tools. In: Proceedings of the 15th international conference on artificial intelligence and law (pp. 69–78)
Guarino A, Lettieri N, Malandrino D, Zaccagnino R (2021) A machine learningbased approach to identify unlawful practices in online terms of service: analysis, implementation and evaluation. Neural Comput Appl 33(24):17569–17587
Hohman F, Head A, Caruana R, DeLine R, Drucker SM (2019) Gamut: A design probe to understand how data scientists understand machine learning models. In: Proceedings of the 2019 chi conference on human factors in computing systems (pp. 1–13)
Japkowicz N, Shah M (2011) Evaluating learning algorithms: a classification perspective. Cambridge University Press, Cambridge
Katz DM (2012) Quantitative legal predictionor- how i learned to stop worrying and start preparing for the data-driven future of the legal services industry. Emory LJ 62:909
Katz DM, Bommarito MJ, Gao S, Arredondo P (2023) Gpt-4 passes the bar exam. Available at SSRN 4389233
Kehl DL, Kessler SA (2017) Algorithms in the criminal justice system: Assessing the use of risk assessments in sentencing
Kim PT (2016) Data-driven discrimination at work. Wm. & Mary L. Rev 58:857
Kim S, Joshi P, Kalsi PS, Taheri P (2018) Crime analysis through machine learning. 2018 IEEE 9th annual information technology, electronics and mobile communication conference (iemcon) (pp. 415–420)
Lettieri N, Altamura A, Malandrino D (2017) The legal macroscope: experimenting with visual legal analytics. Inf Vis 16(4):332–345
Lettieri N, Altamura A, Malandrino D, Punzo V (2017) Agents shaping networks shaping agents: Integrating social network analysis and agent-based modeling in computational crime research. In: Epia conference on artificial intelligence (pp. 15–27)
Lettieri N, Guarino A, Malandrino D (2018) E-science and the law. three experimental platforms for legal analytics. Legal knowledge and information systems - JURIX 2018: The thirty-first annual conference, groningen, the netherlands, 12-14 december 2018. (pp. 71–80). Retrieved from https://doi.org/10.3233/978-1-61499-935-5-71
Lettieri N, Malandrino D, Vicidomini L (2017) By investigation, i mean computation. Trends Organ Crime 20(1–2):31–54
Licklider JC (1960) Man-computer symbiosis. IRE Trans Hum Factors Electron 1(1):4–11
Lin Y-L, Chen T-Y, Yu L-C (2017) Using machine learning to assist crime prevention. 2017 6th IIAI international congress on advanced applied informatics (iiai-aai) (pp. 1029–1030)
McHugh ML (2012) Interrater reliability: the kappa statistic. Biochem Med 22(3):276–282
Meneses-Escobar CA, Castillo-Rodríguez CM, Rodas-Vásquez A (2019) Análisis espacial y temporal del hurto de celulares, pereira, risaralda, año 2018. Revista Logos Ciencia & Tecnología 11(2):167–175
Mohler G, Porter MD (2018) Rotational grid, pai-maximizing crime forecasts. Stat Anal Data Min ASA Data Sci J 11(5):227–236
Nielsen J, Landauer TK (1993) A Mathematical Model of the Finding of Usability Problems. Proceedings of the interact’93 and chi’93 conference on human factors in computing systems (pp. 206–213)
O’Neil C (2016) Weapons of math destruction: How big data increases inequality and threatens democracy. Broadway Books
Ordoñez-Eraso HA, Pardo-Calvache CJ, Cobos-Lozada CA (2020) Detección de tendencias de homicidios en colombia usando machine learning. Revista Facultad de Ingeniería 29(54):e11740–e11740
Pasquale F (2015) The black box society: The secret algorithms that control money and information. Harvard University Press, Cambridge
Ribeiro AL, Araújo OR, Oliveira LB, Inácio M (2022) The executive branch decisionsin brazil: a study of administrative decrees through machine learning and network analysis. PLoS ONE 17(7):e0271741
Richardson R, Schultz JM, Crawford K (2019) Dirty data, bad predictions: how civil rights violations impact police data, predictive policing systems, and justice. NYUL Rev Online 94:15
Rosasco L, De Vito E, Caponnetto A, Piana M, Verri A (2004) Are loss functions all the same? Neural Comput 16(5):1063–1076
Rummens A, Hardyns W, Pauwels L (2017) The use of predictive analysis in spatiotemporal crime forecasting: building and testing a model in an urban context. Appl Geogr 86:255–261
Rzecki K, Plawiak P, Niedźwiecki M, Sośnicki T, Leśkow J, Ciesielski M (2017) Person recognition based on touch screen gestures using computational intelligence methods. Inf Sci 415:70–84
Scaglione A (2011) Reti mafiose: Cosa nostra e camorra: organizzatori criminali a confronto. Reti mafiose, 0–0
Shapiro SS, Wilk MB (1965) An analysis of variance test for normality (Complete Samples). Biometrika 52(3/4):591–611
Shneiderman B (2020) Bridging the gap between ethics and practice: guidelines for reliable, safe, and trustworthy humancentered ai systems. ACM Trans Interact Intell Syst 10(4):1–31
Tashea J (2017) Calculating crime. ABAJ 103:54
Vapnik V (1999) The nature of statistical learning theory. Springer, Berlin
Wachter S, Mittelstadt B (2019) A right to reasonable inferences: re-thinking data protection law in the age of big data and ai. Colum, Bus. L. Rev., p 494
Wang T, Rudin C, Wagner D, Sevieri R (2013) Learning to detect patterns of crime. Joint European conference on machine learning and knowledge discovery in databases (pp. 515–530)
Wasserman S, Faust K (1994) Social network analysis: methods and applications. Cambridge University Press, Cambridge
Weld DS, Bansal G (2019) The challenge of crafting intelligible intelligence. Commun ACM 62(6):70–79
Wenskovitch J, Zhou M, Collins C, Chang R, Dowling M, Endert A, Xu K (2020) Putting the “i’’ in interaction: interactive interfaces personalized to individuals. IEEE Comput Gr Appl 40(3):73–82
Wheeler AP, Steenbeek W (2021) Mapping the risk terrain for crime using machine learning. J Quantit Criminol 37:445–480
Xu W (2019) Toward human-centered ai: a perspective from human-computer interaction. interactions 26(4):42–46
Zheng N-N, Liu Z-Y, Ren P-J, Ma Y-Q, Chen S-T, Yu S-Y, Wang F-Y (2017) Hybrid-augmented intelligence: collaboration and cognition. Front Inf Technol Electron Eng 18(2):153–179
Funding
This study has received no funding.
Author information
Authors and Affiliations
Contributions
Conceptualization: NL; Methodology: AG, RZ, NL; Formal analysis and investigation: NL; Project administration: DM; Software: Alfonso Guarino; Visualization: NL, DM; Method: AG, RZ; Data curation: NL, DM; Writing Original Draft: NL, AG; Writing Reviewing and Editing: NL, DM, RZ; Supervision: DM, NL.
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no conflict of interest.
Ethical approval
This article does not contain any studies with human participants or animals performed by any of the authors.
Informed consent
The research did not involve human participants.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
A Details of crime profile features employed by the online learning model
Here, we provide further details on the features \(f_1\) and \(f_2\) engineered for our AI model concerning the imprisonment months foreseen by the Italian Law for the different crimes.
B Acronyms’ list
Here, we provide the list of acronyms used throughout this paper.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Lettieri, N., Guarino, A., Zaccagnino, R. et al. Keeping judges in the loop: a human–machine collaboration strategy against the blind spots of AI in criminal justice. Soft Comput 27, 11275–11293 (2023). https://doi.org/10.1007/s00500-023-08604-z
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00500-023-08604-z