Abstract. The areas within which the technological revolution set in motion by the AI could more significantly impact with the claims of protection of legal assets, entrusted to criminal law, are basically four: law enforcement and, in particular, predictive policing, where AI systems can make an important contribution to combating, or better yet preventing, the commission of crimes; the possible use of decision-making algorithms to resolve criminal disputes, in order to operate a sort of replacement, or at least of juxtaposition, of the judge-man with the judge-machine; the crime risk assessment entrusted to predictive algorithms, able to draw and re-elaborate enormous quantities of data in order to bring out relationships, coincidences, correlations, that make it possible to profile a person and predict his or her subsequent behavior, even criminal; finally, the possible hypotheses of involvement – as a tool, as an author, or as a victim – of an AI system in the commission of a crime. This work aims to illustrate these areas, indicating problems and prospects connected to the use of AI systems.
SUMMARY: 1. Introduction: survey limits and purposes. – 2. What do we mean by artificial intelligence? – 3. First research lead – AI and law enforcement. – 3.1. RoboCop: from science fiction to reality? – 3.2. Artificial intelligence systems and predictive policing. – 3.2.1. Hotspots detection systems. – 3.2.2. Crime linking systems. – 3.2.3. Concluding remarks on predictive policing. – 4. Second research lead – AI and judicial decision: the judge-machine? – 5. Third research lead – AI and crime risk assessment: predictive algorithms. – 5.1. Introductory remarks. – 5.2. The “actuarial” crime risk assessment. – 5.3. The use of predictive algorithms in the United States. – 5.3.1. Psa – Public Safety Assessment. – 5.3.2. COMPAS – Correctional Offender Management Profiling for Alternative Sanctions. – 5.3.2.1. In particular, the Loomis case and the controversial use of COMPAS in sentencing. – 5.4. Concluding remarks. – 6. Fourth research lead – AI and crime: possible hypotheses of involvement – as a tool, as an author, or as a victim – of an AI system in the commission of a crime. – 6.1. Introductory remarks. – 6.2. The AI system as a tool to commit crimes. – 6.3. The AI system as the perpetrator of the crime: machina delinquere potest? – 6.3.1. Between loss of human responsibility and machine accountability. – 6.3.2. Does the boundary between machina and human falter? – 6.3.3. An “inhuman” guilt? – 6.3.4. What penalties for AI systems? – 6.4. The AI system as a crime victim. – 7. What future awaits us?
To read the Reflection, click on “open file”.
Quest'opera è distribuita con Licenza Creative Commons Attribuzione - Non commerciale 4.0 Internazionale
