Search

N/A

The principles of transparency and explainability are landmarks of the current EU approach to artificial intelligence. Both are invoked in the policy guidelines as values governing algorithmic decision-making, while providing rationales for existing normative provisions, on information duties, access rights and control powers. This contribution addresses the debate on transparency and explainability from the EU consumer market perspective. The consumers’ position relative to algorithmic decision-making is considered, and their risks concerning mass surveillance, exploitation, and manipulation are discussed. The concept of algorithmic opacity is analyzed, distinguishing technology-based opacity that is intrinsic to design choices, from relational opacity toward users. The response of EU law is then considered. The emerging approach to algorithmic transparency and explainability is connected to the broader regulatory goals concerning transparency in consumer markets. It is argued that EU law focuses on adequate information being provided to lay consumers (exoteric transparency), rather than on understandability to experts (esoteric transparency). A discussion follows on the benefits of transparency, on its costs, and on the extent to which transparency can be implemented without affecting performance. Finally, the merits of a transparency-based regulation of algorithms are discussed and insights are provided on regulating transparency and explainability within the EU law paradigm.

Freie, öffentlichen Meinungsbildung ist das Herzstück der Demokratie. Doch digitale Kommunikation und datengetriebene Kuratierung von Inhalten verändern das der Demokratie eigene Konzept von Öffentlichkeit und erfordern neue gesetzliche Rahmenbedingungen. In diesem Sammelband führen Expert:innen der Rechts- und Politikwissenschaften, der Soziologie und Datenwissenschaft in die Materie ein und weisen Wege zur Stärkung der Demokratie in der Digitalisierung.

New technologies have always challenged the social, economic, legal, and ideological status quo. Constitutional law is no less impacted by such technologically driven transformations, as the state must formulate a legal response to new technologies and their market applications, as well as the state’s own use of new technology. In particular, the development of data collection, data mining, and algorithmic analysis by public and private actors present unique challenges to public law at the doctrinal as well as the theoretical level. This collection, aimed at legal scholars and practitioners, describes the constitutional challenges created by the algorithmic society. It offers an important synthesis of the state of play in law and technology studies, addressing the challenges for fundamental rights and democracy, the role of policy and regulation, and the responsibilities of private actors. This title is also available as Open Access on Cambridge Core.

Preface to the volume ‘WOA 2021 – 22nd Workshop “From Objects to Agents”’

We present a logical analysis of influence and control over the actions of others, and address consequential causal and normative responsibilities. We first account for the way in which influence can be exercised over the behaviour of autonomous agents. On this basis we determine the conditions under which influence leads to control on the implementation of positive and negative values. We finally define notions of causal and normative responsibility for the action of others. Our logical framework is based on STIT logic and is complemented with a series of examples illustrating the application. Our analysis applies to interactions between humans as well as to those involving autonomous artificial agents.

Da qualche anno, il governo cinese sta costruendo un sistema per la raccolta, conservazione e uso di grandi masse di dati digitali che riguardano il comportamento delle persone fisiche e giuridiche in ambito economico, finanziario, sociale, civile e altri ancora. Tali dati saranno usati per valutare il comportamento dei soggetti attribuendo incentivi per le condotte “affidabili”, “degne di fiducia” o “virtuose”, e limitazioni per quelle “non affidabili”, “non degne di fiducia”, “non virtuose”, secondo parametri giuridici, di moralità sociale o proposti dal potere politico. L’attribuzione del credito sociale e la determinazione delle sue conseguenze risultano da un complesso sistema socio-tecnico (combinazione di persone, norme e tecnologie) nel quale le tecnologie dell’informatica, e in particolare quelle dell’intelligenza artificiale, giocano un ruolo fondamentale. Il presente contributo si propone di introdurre le caratteristiche, la struttura, le finalità esplicite e i principali limiti del sistema di credito sociale, illustrando brevemente le potenzialità, ancora non del tutto attualizzate, dell’uso delle tecniche di intelligenza artificiale nelle diverse applicazioni relative alla sorveglianza, alla valutazione e alla sanzione. In particolare, si porrà l’attenzione sulle caratteristiche del modello sanzionatorio del sistema di credito sociale. Esso è rivolto a realizzare importanti esigenze sociali ma, d’altro lato, è suscettibile di incidere pesantemente sulle libertà dei cittadini e sullo sviluppo delle dinamiche sociali. La sua valutazione richiede un’analisi accurata, che include il contesto socio-culturale in cui esso si colloca così come le tecnologie di cui si avvale. Over the last few years, the Chinese government has built a system for the collection, conservation and utilization of large masses of Big Data concerning the behavior of natural and legal persons in several (economic, financial, social, civil, etc.) fields. All these data are to be used to assess the behavior of the individuals, attributing incentives for “reliable”, “trustworthy” or “virtuous” conducts, and limitations for “unreliable”, “untrustworthy” or “unvirtuous” conducts, according to parameters coming from the legal sphere, from the social morality or proposed by the political power. The attribution of social credit and the determination of its consequences result from a complex socio-technical system (a combination of persons, rules and technologies) in which information (particularly AI) technologies are paramount. This paper aims to present the characteristics, the structure, the explicit goals and the main limits of the social credit system, briefly illustrating the potential of AI technologies in the applications concerning surveillance, evaluation and sanction. In particular, it will highlight the characteristics of the sanctioning model of the social credit system, which aims to realize important social goals, but which can also heavily affect the freedom of citizens and a number of social dynamics. Thus, the evaluation of this system requires a precise analysis, which includes its socio-cultural context and the technologies used by it.

The study addresses the regulation of targeted and behavioural advertising in the context of digital services. Marketing methods and technologies deployed in behavioural and target advertising are presented. The EU law on consent to the processing of personal data is analysed, in connection with advertising practices. Ways of improving the quality of consent are discussed as well as ways of restricting its scope as a legal basis for the processing of personal data. This study is commissioned by the European Parliament’s Policy Department for Citizens’ Rights and Constitutional Affairs at the request of the JURI Committee.

Il contributo esamina l'articolo 22 GDPR, in merito al processo decisionale automatizzato relativo alle persone fisiche, compresa la profilazione, approfondendo del diritto alla spiegazione delle decisioni automatizzate.

Criminal liability for acts committed by AI systems has recently become a hot legal topic. This paper includes three different contributions. The first contribution is an analysis of the extent to which an AI system can satisfy the requirements for criminal liability: accomplishing an actus reus, having the corresponding mens rea, possessing the cognitive capacities needed for responsibility. The second contribution is a discussion of criminal activity accomplished by an AI entity, with reference to a recent case involving an online bot, the Random Darknet Shopper. This discussion will provide the context for the analysis of commonalities and differences between criminal activities by humans and by artificial systems. The third contribution concerns the evaluation of different ways of addressing criminal activities by AI systems in a regulatory perspective.

Il contributo esamina il concetto di intelligenza artificiale e le caratteristiche distintive dei modelli logici rispetto all'apprendimento automatico. Analizza le opportunità e i rischi dell'IA nel nuovo millennio, con particolare attenzione al fenomeno dei big data e ai rischi legati alla discriminazione e alla privacy, identificando principi, valori e norme giuridiche necessari per costruire un quadro etico e giuridico dell'intelligenza artificiale.

This paper provides a formal model for the burden of persuasion in dialogues, and in particular, in legal proceedings. The model shows how an allocation of the burden of persuasion may induce single outcomes in dialectical contexts in which, without such an allocation, the status of conflicting arguments would remain undecided. Our approach is based on a two-stage labelling. The first-stage labelling determines what arguments are accepted, rejected or undecided, regardless of the allocation of burden. The second-stage labelling revises the dialectical status of first-stage undecided arguments, according to burdens of persuasion. The labelling is finally extended in such a way as to obtain a complete labelling. Our model combines two ideas that have emerged in the debate on the burden of persuasion: the idea that the burden of persuasion determines the solution of conflicts between arguments, and the idea that its satisfaction depends on the dialectical status of the arguments concerned. Our approach also addresses inversions of the burden of persuasion, namely, cases in which the burden of persuasion over an argument does not extend to its subarguments., In Proceedings ICLP 2020, arXiv:2009.09158

Il contributo esamina il ruolo e l’impatto della profilazione degli individui, e in particolare degli utenti delle piattaforme digitali, sui meccanismi sociali, economici e politici che regolano il funzionamento della Rete, del mercato e più in generale della sfera pubblica, con attenzione al fenomeno delle c.d. fake news. La profilazione e i meccanismi di sorveglianza sempre più persistenti e pervasivi, messi in campo grazie alle più recenti tecniche dell’intelligenza artificiale, sono analizzati con particolare rifermento alle capacità di classificazione, valutazione, influenza e manipolazione degli individui. Il contributo propone quindi un’analisi delle nuove frontiere dell’ecosistema della sorveglianza nella sfera pubblica, con particolare riferimento all’opinione pubblica e alle preferenze di voto, fino a ricomprendere fenomeni quali ad esempio il c.d. Sistema di credito Sociale Cinese. Infine si esamina il nesso tra la regolazione dell’intelligenza artificiale e le principali sfide tecnologiche dei prossimi anni. The paper aims at analysing the role and the impact of user profiling (particularly on the Internet platforms) on the social, economical and political mechanisms of functioning of the Internet, the market, and, more broadly, the public sphere. In this paper, particular attention to the issue of the so-called fake news is paid. In this sense, the increasingly persistent and pervasive user profiling and surveillance mechanisms developed due to the more recent artificial intelligence techniques are analysed by looking at the issues of the scope of classification, evaluation, influence, and manipulation of individuals. As a consequence, the paper explores the new frontiers of the surveillance ecosystem in the public sphere, with particular reference to public opinion and voting preferences, including phenomena such as the so-called Chinese Social Credit System. Finally, the link between the regulation of artificial intelligence and the main technological challenges of the upcoming years is analysed.

Preface to the Proceedings of the 21st Workshop “From Objects to Agents”

After introducing the concept and the methods of artificial intelligence, we shall consider the prospects of machine learning technologies, as applied to big data. We shall discuss the opportunities and risks of these technological developments. In particular we shall consider how AI on the one hand further expands the prevailing powers of big companies and governments, but on the other hand, can also support individuals and their organisations. AI can indeed strengthen the countervailing powers of civil society, so as to facilitate in particular the enforcement of the law. As an example, we shall present the Claudette project, which uses AI techniques to identify unfair and unlawful content in online contracts and privacy policies.

One of the contested issues in this debate relates to similarities and differences between scoring systems in China and in the West – how unique is China? In this post, we will try to reconcile the different perspectives, arguing that both commonalities and differences exist, depending on the adopted level of abstraction. Thus, we shall zoom in the Chines Social Credit System (SCS), examining the features it shares with other systems and point to related issues: it is a scoring system, it is formal, it is ICT based, it is surveillance based, it is opaque and unaccountable. This enables us to distinguish commonalities and differences.