Technological progress could constitute a huge benefit for law enforcement: greater efficiency, effectiveness and speed of operations as well as more precise risk analyses, including the discovery of unexpected correlations, which could feed nourish profiles. A number of new tools entail new scenarios for information gathering, as well as the monitoring, profiling and prediction of individual behaviours, thus allegedly facilitating crime prevention: algorithms, artificial intelligence, machine learning and data mining. Law enforcement authorities have already embraced the assumed benefits of big data. However, there is a great need for an in-depth debate about the appropriateness of using algorithms in machine-learning techniques in criminal justice, assessing how the substance of legal protection may be weakened. Given that big data, automation and artificial intelligence remain largely under-regulated, the extent to which data-driven surveillance societies could erode core criminal law principles such as reasonable suspicion and the presumption of innocence, ultimately depends on the design of the surveillance infrastructures. This contribution first addresses the so-called rise of the algorithmic society and the use of automated technologies in criminal justice to assess whether and how the gathering, analysis and deployment of big data are changing law enforcement activities. It then examines the actual or potential transformation of core principles of criminal law and whether the substance of legal protection may be weakened in a ‘data-driven society’.

Technologies have always challenged, if not disrupted, the social, economic legal, and to an extent, the ideological status quo. Such transformations impact constitutional law, as the State formulates its legal response to the new technologies being developed and applied by the market, and as it considers its own use of the technologies. The development of data collection, mining, and algorithmic analysis, resulting in predictive profiling – with or without the subsequent potential manipulation of attitudes and behaviors of users – presents unique challenges to constitutional law at the doctrinal as well as theoretical levels.

Statutory interpretation is a problem that is of crucial importance for legal practice and theory, political discussions, ethical issues, and public information. A substantial majority of the US Supreme Court’s case load involves statutory construction, nearly two-thirds of its docket by one recent estimate, and in the years ahead, courts will be asked to construe the meaning of thousands of sections of legislation. However, as Katzmann (2014, 3–10) emphasized, the interpretation of legal texts affects not only daily rulings in the courts at all levels, but even what have become political issues vital to the legal system such as televised confirmation hearings for US Supreme Court nominees. Statutes affect all aspects of our daily lives, including the most pressing public policy issues at a given time. However, although ideally the language of the statute should be clear, the texts passed by legislative bodies, such as the US Congress, can be vague, ambiguous, structurally complex in expression, or even apparently logically inconsistent. Fundamental values of our societies, emerging in the controversies on the issues of freedom of speech, abortion, marriage, or self-defense, are primarily debated as matters of interpretation.

In our previous chapter on ambiguity, we underscored how the logical form of a legal statement does not provide the proposition expressed. The meaning of an utterance – the product of a verbal act performed in a specific context (Leech 1983, 14) – cannot be the simple output of a decoding process (Sperber and Wilson 1995, 182; Recanati 2003, 56), or “semantic interpretation” (Leech 1983, 5). The logical form (also called “semantic representation”) that can be recovered through the mere decoding of an utterance through the application of the rules of grammar (Sperber and Wilson 1995, 9–10) does not deliver complete propositions, but only “semantic schemata” (Recanati 2003, 56).

For many readers the paradigm of argumentation in a legal setting is that of the trial, an instance of the type of dialogue called the persuasion dialogue or critical discussion in the argumentation literature (van Eemeren and Grootendorst, 1992; Walton, 1999; Prakken, 2009). This setting is an adversarial one. For example, in a criminal case the prosecution’s role is not only to support its claim that the defendant committed the crime by bringing forward evidence, but also to attack the arguments of the other side, defend its own arguments against these attacks, and prove its own claim to the standard required, that of beyond reasonable doubt in the common law system. With this paradigm in mind, it is easy to jump to the generalization that legal argumentation is a kind of persuasion dialogue in which each side is trying to persuade the trier of fact to accept its view of the matter (Feteris, 1999, 171–174; Kloosterhuis, 2013).

In our previous chapters, we underscored the role of interpretative arguments in the justification of the passage from a legal text to a legal rule (Hage, 1996, 214; Tarello, 1980), where the latter is intended as a normative premise under which an individual case can be “subsumed” or classified (Moreso & Chilovi, 2018). Legal interpretation can be compared to the common understanding and processing of utterances in ordinary conversation (Smolka & Pirker, 2016), in which semantic content is only a vehicle for getting to the “speaker’s meaning,” i.e., what is communicated – a richer content to which semantic “meaning and obvious background assumptions have both contributed” (Soames, 2008, 411; see also Butler, 2016; Carston, 2013; Horn, 1995; Miller, 1990).