Prof. Maristella Amisano is Professor of Criminal Law at the University of Calabria, Italy, where she teaches Criminal Law II and Criminal Law of the Economy. She graduated in Law from the University of Turin in 1994–1995 with a thesis in Philosophy of Law on paternalism in the informational-normative context of complex society, receiving a final grade of 106/110 with merit. She subsequently undertook legal training in criminal law and continued her professional and academic development through collaboration with the Italian State Attorney’s Office in Turin and with the law firm of Prof. Avv. Marcello Gallo. She qualified as a lawyer in 1999 and has been registered with the Bar of Turin since that year, while since 2006 she has also been registered with the Bar of the Italian Supreme Court (Corte di Cassazione). In 1999, she obtained a Master’s degree in Criminology from Adelphi University in New York. Prof. Amisano has held numerous academic and teaching positions throughout her career. From 2002 to 2006, she was Professor of Criminal Law at the Faculty of Political Sciences of the University of Calabria. In 2006, she successfully passed the national competition for Associate Professor in the IUS/17 sector and took up a professorship at the Faculty of Economics of the University of Calabria. Since 2010, she has served as a full-time professor and has taught a broad range of subjects, including Criminal Law I and II, Criminal Law of the Economy, Criminal Labour Law, and Criminal Procedure. She is also actively involved in postgraduate education, including the School of Specialization for Legal Professions (SSPL Fausto Gullo), where she has taught criminal law, coordinated seminars, supervised examinations and supported trainees. She is a member of the Scientific Committee of the School of Specialization and has served as a tutor and supervisor for doctoral students. Her academic and professional activities extend beyond teaching. She has served as a member of examination commissions for the Italian legal profession, as a scientific committee member of doctoral and postgraduate programs, and as a referee for academic and research evaluation bodies. Since 2020, she has been a member of the editorial direction of the journal La Giustizia Penale and is also a member of the Scientific Committee of the journal Teoria e Prassi del Diritto. Since 2022, she has served as Delegate for Research of the Director of the Department of Business and Legal Sciences at the University of Calabria. She is also a founding member of the Alimena Institute of Criminal Studies, an interdisciplinary research centre of the University of Calabria, and since 2026 has been a member of the Zaleuco Interdepartmental Centre of the University of Calabria, dedicated to interdisciplinary research in legal and social sciences. Prof. Amisano has developed significant international academic collaborations, particularly with Brazil. She is a member of CONPEDI (Conselho Nacional de Pesquisa e Pós-Graduação em Direito) and Scientific Director of Lar do Consumidor, a Brazilian legal institution operating with the support of several Brazilian public institutions. She has collaborated in research with scholars and universities including the Federal University of Minas Gerais (UFMG), the Federal University of Uberlândia, the State University of Belo Horizonte and the Pontifical Catholic University of Minas Gerais. She has delivered numerous lectures and seminars in Italian, Portuguese and English at Brazilian universities, including UFMG and Universidade Milton Campos, addressing topics such as Italian criminal law, comparative criminal law, corruption, anti-corruption strategies, criminal liability, sexual offences, organised crime and cooperation with justice. Her international academic experience includes participation as an Italian representative at the World Congress of the International Association of Penal Law (AIDP) in Rio de Janeiro in 2014. She has also served as a speaker at international conferences in Brazil and Europe, including events at the Federal University of Minas Gerais, the University of Timișoara and other international academic institutions. Her conference contributions have addressed a wide range of contemporary criminal-law issues, including corruption, money laundering, criminal liability, victims of crime, confiscation, environmental criminal law, media and criminal justice, gender-based violence, femicide and emerging challenges to criminal law. Her research interests cover a broad spectrum of criminal-law and criminological issues. She has been responsible for numerous university research projects addressing corruption, comparative Italian and Brazilian criminal law, sentencing, intent and negligence, dangerousness, migration and cultural defence, terrorism and criminal law, media influence on criminal justice, torture, punishment, fundamental rights during emergencies, artificial intelligence and neurosciences, protection of truth, environmental criminal law, judicial discretion in sentencing, and gender-based violence. In 2021, she led a research project entitled “Artificial Intelligence and Neurosciences: New Instruments of Criminal Law?”, reflecting her longstanding interest in the legal challenges posed by emerging technologies. Within the Italian National Recovery and Resilience Plan (PNRR) and the Tech4You innovation ecosystem, Prof. Amisano is responsible for two pilot projects focusing on the integration of legal instruments for environmental sustainability, circular economy regulation and the reuse of secondary raw materials. She has also participated in national research projects concerning drug policy and the legalisation of soft drugs, as well as the SERICS project (Security and Rights in Cyberspace). In recent years, her research and teaching have increasingly focused on contemporary challenges to criminal law arising from technological, environmental and social transformations. In 2024, she delivered a doctoral lecture addressing autonomous vehicles, ChatGPT and the metaverse. In 2025, she addressed the prevention of environmental crime and the relationship between criminal and administrative law. In 2026, she presented the topic “Culpa in Algorithmo”, examining the possibility of criminal liability for artificial intelligence within the Italian legal system. Her recent academic activities have also addressed the challenges of modernity to criminal law, criminal deviance, the criminal mind, organised crime, gender-based violence and femicide. Prof. Amisano has extensive experience in academic research, legal education, comparative criminal law and international cooperation. She has authored and presented research across several areas of contemporary criminal law and continues to contribute to the development of Italian and comparative legal scholarship through teaching, research, academic supervision and international collaboration. She has an excellent command of Portuguese, a good command of English, and a working understanding of French and Spanish.

CULPA IN ALGORITHMO: CAN CRIMINAL LIABILITY FOR ARTIFICIAL INTELLIGENCE BE CONCEIVED WITHIN THE ITALIAN LEGAL SYSTEM? Artificial intelligence (AI) has become an increasingly indispensable component of contemporary social life and, as such, calls for a legal framework capable of regulating its use consistently with fundamental rights. Social change is necessarily reflected in law, which must adapt to and govern the application of technological innovation. Because artificial intelligence is not confined to a single sector but increasingly permeates the legal and social order as a whole, it cannot be adequately examined through the lens of a single discipline. This is particularly true of criminal law, whose protected interests are ultimately rooted in society itself and which must therefore remain receptive to technological development rather than retreating into anachronistic isolation. The increasing deployment of artificial intelligence may give rise to new and complex forms of legal liability, including, potentially, criminal liability. The central question of this inquiry may be stated in deceptively simple terms: when artificial intelligence causes harm or otherwise produces an unlawful outcome, who should be held responsible? Liability, in its broadest sense, presupposes the capacity to answer for conduct and to bear the legal consequences arising from it. Although this concept is common to the legal order as a whole, the present analysis focuses specifically on its criminal-law dimension. Identifying the subject to whom a criminally relevant fact may be attributed therefore provides the organising principle of the analysis. Italian criminal law predicates liability upon a legally defined offence: conduct corresponding to an offence established by statute and accompanied by the requisite subjective element. For this reason, criminal law remains distinct from regulatory frameworks such as the European Union Artificial Intelligence Act, which primarily establishes regulatory requirements, obligations, and standards rather than autonomous categories of criminal responsibility. Criminal law is not concerned, in the first instance, with whether a particular use of artificial intelligence is ethically acceptable; rather, it asks whether a legally relevant fact has occurred that satisfies the statutory elements of an offence and whether the conduct was committed intentionally (dolo), negligently (colpa), or with preterintenzione, that is, where the agent intended a less serious result but caused a more serious one. Accordingly, there can be no generic form of criminal liability for “artificial intelligence” as such. Criminal responsibility must be established in relation to a specific offence and must encompass both its objective (elemento oggettivo) and subjective (elemento soggettivo) elements. Artificial intelligence enters contemporary criminal-law discourse principally in two ways. The first, and currently the most common, concerns offences committed by natural persons who employ AI as an instrument of criminal conduct. For example, Article 612-quater of the Italian Criminal Code addresses the unlawful dissemination of sexually explicit images or videos generated or manipulated through technological means, including through the use of artificial intelligence, where the statutory requirements are satisfied. A similar instrumental paradigm may be observed in fraud schemes targeting elderly victims, in which offenders employ AI-generated or manipulated voices and images to create fictitious emergencies and induce victims to transfer money. Comparable techniques may also be used in so-called romance scams, where AI-generated content is employed to impersonate public figures or other individuals. In such cases, artificial intelligence functions as a technologically sophisticated instrument placed at the service of an unlawful purpose. Its legal function is, in principle, comparable to that of any other instrument employed in the commission of an offence—for example, a weapon or other object used to perpetrate a homicide. Criminal responsibility therefore remains attributable to the natural person who intentionally or negligently employs the technology, provided that the relevant statutory requirements are satisfied. A distinct set of issues arises in the context of autonomous vehicles. If a self-driving vehicle is deliberately employed to transport narcotics, criminal responsibility will ordinarily remain attributable to the individual who intentionally uses the vehicle for that unlawful purpose. If, by contrast, an autonomous vehicle strikes and injures or kills a pedestrian, the question becomes substantially more complex. Liability must then be assessed by identifying the natural person whose conduct may legally be regarded as causally and normatively connected to the malfunction or harmful outcome, which may potentially include a programmer, manufacturer, operator, or another responsible actor. In such circumstances, however, the ordinary principles governing causation, foreseeability, negligence, and personal culpability remain applicable. An even more difficult scenario arises where an individual suffers serious harm following prolonged interaction with an artificial intelligence system. The death of Adam Raine, a sixteen-year-old who died by suicide following an extended series of interactions with an AI system, illustrates the legal difficulties that may arise in such circumstances. Although the legal debate surrounding such cases may primarily concern the potential civil liability of developers, providers, or distributors, the criminal-law perspective presents a distinct challenge. Under the present legal framework, criminal responsibility would most plausibly have to be attributed to a natural person whose conduct satisfies the elements of a specific offence—for example, where negligent design, inadequate safeguards, or a failure to comply with a legally relevant duty can be established. Once again, criminal liability would ultimately depend upon identifying a human actor whose conduct satisfies the requirements of the applicable criminal provision. The progressive development of machine-learning systems, however, may profoundly complicate this traditional model of attribution. As programmers, developers, operators, and other human actors become increasingly removed from the concrete decisions ultimately generated by autonomous systems, the causal and, above all, subjective nexus between a natural person and the resulting harm may become increasingly attenuated. In sufficiently complex cases, it may become difficult—or perhaps impossible—to identify an individual whose conduct can satisfy all the requirements necessary for criminal responsibility. At that point, a question that currently appears largely theoretical could acquire concrete legal significance: could criminal responsibility ever be attributed directly to an artificial intelligence system itself? Because criminal law represents one of the most significant forms of State coercive power, the question of AI liability must necessarily be examined within the constitutional and doctrinal framework of each individual legal system. The present analysis therefore focuses specifically on Italian law, while also identifying difficulties that may arise more broadly within civil-law jurisdictions. Article 27 of the Italian Constitution provides that “criminal liability is personal.” Traditionally, this constitutional principle was understood as limiting criminal responsibility to natural persons. Such an interpretation was reinforced by the apparent incompatibility between criminal punishment and non-human entities, particularly in light of the traditional importance of imprisonment and other sanctions designed for human offenders. It also found expression in the longstanding maxim societas delinquere non potest, according to which a legal entity could not itself be the bearer of criminal responsibility. Contemporary constitutional and criminal-law scholarship, however, has increasingly questioned an exclusively anthropocentric interpretation of Article 27. From this perspective, the requirement that criminal liability be “personal” does not necessarily imply that the liable subject must, in every conceivable circumstance, be a natural person. Rather, it may be understood as requiring a sufficiently demonstrable psychological and normative connection between the subject to whom the offence is attributed and the criminally relevant conduct, thereby excluding forms of strict or purely objective liability based solely upon causal contribution to a prohibited result. This doctrinal development makes it conceptually possible to consider whether forms of legal responsibility might, under certain conditions, be attributed to entities other than natural persons. The question nevertheless becomes considerably more difficult when the entity in question is not a legal person, but an artificial system capable of autonomous decision-making. A highly autonomous and cognitively sophisticated artificial intelligence could, in the future, require the legal system to confront the possibility of a corresponding form of responsibility. Even if Article 27 were interpreted as leaving room for forms of responsibility extending beyond natural persons, at least two fundamental obstacles would remain. The first concerns the identification of a sufficiently meaningful subjective or psychological element within an artificial system. The second concerns the construction of a sanctioning regime capable of satisfying the constitutional purposes of punishment, including the principle of rehabilitation recognised by Article 27 of the Italian Constitution. Neither difficulty is necessarily conceptually insurmountable. Italian Legislative Decree No. 231 of 2001 provides an important legal and conceptual reference point. The decree establishes a system of liability for entities in connection with certain criminal offences, incorporating both objective and subjective attribution requirements and providing for a structured catalogue of sanctions. Although this regime is formally characterised as responsabilità amministrativa da reato, rather than criminal liability in the strict doctrinal sense, its substantive proximity to criminal responsibility has led some scholars to question whether the distinction is, at least in part, one of legal classification or “labelling.” Regardless of the doctrinal debate surrounding its precise legal nature, Legislative Decree No. 231/2001 demonstrates that the Italian legal order is already capable of attributing a legally autonomous form of responsibility to an entity other than a natural person. It also demonstrates that the legal system can construct mechanisms of attribution and sanctions directed towards entities that do not possess the same characteristics as individual human offenders. Against this background, Italian law may already contain some of the conceptual instruments that could, should technological and social developments make such a step necessary, contribute to the construction of a framework for responsibility directly attributable to artificial intelligence. Nevertheless, the transition from liability of human beings and legally recognised entities to criminal responsibility of autonomous artificial systems would require profound reconsideration of fundamental concepts, including culpability, intentionality, negligence, causation, legal personality, punishment, and rehabilitation. The transformation brought about by artificial intelligence will inevitably generate corresponding challenges for the legal system, and criminal law can be no exception. The issue is therefore no longer simply whether one should be “for” or “against” criminal liability for machines. The deeper question is how the legal order can preserve the principles of legality, personal culpability, accountability, and fundamental rights in a world increasingly structured, mediated, and governed by autonomous computational systems. The Italian legal system provides a particularly valuable framework within which to explore this question, not only because of the constitutional centrality of personal criminal responsibility, but also because its experience with entity liability demonstrates that the boundaries between human agency, institutional responsibility, and technologically mediated conduct are capable of evolving. The ultimate challenge may therefore be to determine whether culpa in algoritmo can ever become a legally meaningful concept without undermining the fundamental principles upon which criminal law itself is founded. References Basile, F., “I.A. e diritto penale: quattro possibili percorsi di indagine,” Diritto penale contemporaneo, 29 September 2019. Boden, M., Artificial Intelligence: A Very Short Introduction, Oxford University Press, 2018. Cappellini, A., “Machina delinquere non potest,” disCrimen, 27 March 2019. Canzio, V. G., “I.A., algoritmi e giustizia penale,” Sistema Penale, 8 January 2021. Quattrocolo, V. S., Artificial Intelligence, Computational Modelling and Criminal Proceedings, New York, 2020. Wiener, N., Introduzione alla cibernetica. L’uso umano degli esseri umani, Turin, 2012.