LAST-JD - All the cycles

Social robots are increasingly becoming part of our reality. They are examples of embodied-AI, able to develop social competences, to appropriately use natural signals and to simulate feelings and empathy. In order to directly interact with the user and to pursue everyday tasks – from education to health, from entertainment to the care of fragile peoples – they are designed so as to elicit human tendency to anthropomorphism. In spite of that, they are just objects, unable to perceive their own existence and to pone their own goals. This might lead to a completely new form of manipulation, potentially risky for the psychological, physical and economical integrity of the humans involved and for the authenticity of their will. The project is intended to analyse the phenomenon of deception in social robotics, investigating how to tackle it from a technological, ethical and legal point of view. The main aim is to ensure a technological development that is effectively focused on human beings and on the timely protection of their physical and mental integrity.

The thesis will focus on the issues brought by the implementation of AI and Big Data technologies in the healthcare context, from a data protection law perspective, and whether self and co-regulation mechanisms could be helpful to face these problems.
Recently, the application of AI technologies is spreading in several medical applications, from robotic surgery to smart pills or mobile health. Nowadays, surgical robots can assist the human surgeon accomplishing very simple tasks during the operations, such as suturing tissues. ML software, on the other hand, can dig in big amount of data trying to predict the occurrence or the evolution of disease.
All these technologies work thanks to the collection and the elaboration of big amount of data gathered from different sources, i.e., they work on Big Data analytics technologies.
Especially in the healthcare, such data are also personal data gathered directly from patients or from hospitals records or even from new and innovative sources, for example wearable devices or mobile apps.
Without questioning the great opportunities that such AI technologies can bring to healthcare processes and clinical research, the elaboration of Big Data in healthcare must be object of further investigation. Several risk could stem indeed for individuals and patients, not only in relation to privacy violation but also to discriminations and unfair decision made by the AI system.
Several factors could impact the accuracy of an AI systems, from biases during data collection or dataset testing operations, to black boxes algorithm that can result in unexplainable decisions made by the AI system.
EU data protection law introduce several requirements meant to face risks of discrimination and un-fair decision adopted by AI systems. However, also other initiative in the EU (e.g., the recent AI proposal) are trying to provide a secure environment where AI systems can be developed safely.
Both the GDPR and other connected regulations introduce the opportunity to rely on self and Co-regulation mechanisms in order to give more practical enforcement to technical requirements in complex environment, such AI and healthcare. Such instruments, such as Codes of conduct, certification mechanism, standardisation, etc. are meant to provide flexibility to rules’ application, but also to help private and public institutions demonstrate compliance with legal requirements.
It should be bear in mind that self and co-regulation mechanisms are also subject to several short-comings, from legitimacy and constitutional issues to difficulty to grant proper enforcement.
In conclusion, the main goal of the thesis is to understand to what extent such self and co-regulation mechanisms could be implemented in healthcare, for facing the mentioned above issues brought by the application of AI technologies and Big Data analytics.

Nowadays, thanks to the advances in medical research and clinical treatments in general, biobanks are becoming essential for the development of precise medicine and the healthcare system. Indeed, biobanks are sometimes considered the bridge from reactive to preventive medicine. However, alongside with enormous advantages, biobanks pose ethical threads and legal issues. Considering that biobanks allow the storage, preservation and future use of samples and data for research project unpredictable at the time of their collection, there is a strong need to protect the patient’s rights and interests, while at the same time pursue the legitimate interests of society at large.
However, biobanks lack a uniform and comprehensive regulation, at least at a European level, thus resulting in a fragmentation of different norms applicable, which makes it even more difficult to share samples and information among research institutions and properly benefit worldwide from the advances in the healthcare sector.
My research project will then focus on the ethical and legal implication of biobanks for the storage of biological samples, health data and microbiome samples for research purposes. I will also analyse the practical implication of the theoretical analysis conducted, in order to develop a biobank which will be implemented by the Istituto Ortopedico Rizzoli.

Ludovica’s research project focuses on European Open Science policies. Open Science can be generally defined as the new paradigm of science, which aims to foster the openness of every phase of the scientific research process, from data collection to the dissemination of scientific results, within the scientific community and externally, towards society. The Open Science purpose is not only the openness of science: the goal is to support good science, in a broad sense, based on communication, cooperation, minimising waste of effort, ensuring transparency and taking advantage of the most innovative technologies. In other words, Open Science is the representation of a new way of doing science that aims to take full advantage of the use of new technologies, cloud computing, high performance computing, and innovative Big Data analytics techniques. In recent years, European institutions seem to have taken a clear stance in supporting and promoting the Open Science paradigm in Europe: projects such as the European Open Science Cloud confirm the relevance of this matter at European level. In the emergence of this new paradigm a several challenges arise, which the research project intends to investigate: challenges related to what is meant by Open Science; and legal challenges, related to the framework of provisions that are involved in this context. Regarding the legal issues, first, the focus is on the link between Open Science projects and the European Open Data framework (with particular reference to the Open Data Directive). Second, the project investigates issues related to data protection and the
derogatory European rules on the processing of personal data for scientific research
purposes. The research project also includes a case study, represented by the High Performance Computing platform of the University of Luxembourg (ULHPC), used for scientific
research purposes. The analysis of the formula Open Science and the study of the legal issues in the field of Open Data and data protection, lead to some final considerations on the governance of scientific research in Europe in relation to the implementation of Open Science projects.

The focus of my research is the issue of “hate speech” in the context of speech regulation within the algorithmic society. Against the backdrop of the rapidly evolving digital environment, the harms connected to such a phenomenon are greatly amplified not only because online content is inherently capable of spreading worldwide, but also because of the specific characters of Internet communication, which can foster extremization of thought. Moreover, today’s online hate speech is strictly interconnected with the phenomenon of so-called ‘fake news’ and disinformation. To face these issues, public institutions are turning more and more to private Internet intermediaries with the purpose of coercing them into proactively fighting against the dissemination of unlawful and harmful content across the Internet. My PhD project aims at addressing the following questions: how should online hate speech governance address these digital private actors? How can the law foster hate speech moderation whilst guaranteeing users’ constitutional rights and what are the risks connected to the private (algorithmic) moderation practices of online platforms?
The work will be structured in three work packages. The first work package will be dedicated to a theoretical inquiry of the rationales justifying the existence of regulatory intervention to fight hate speech, considered as a threat to the well-being of a democratic society and to the fundamental rights of targeted natural persons. The second work package will focus on the increasing role of Internet intermediaries in the fight against illegal and harmful content, a role which is signified by the recent development of instances of “new-school speech regulation” (J. Balkin). In this sense, the research will also address the technical means employed by intermediaries to moderate user-generated third-party content: most notably, the work will reflect on the growing reliance on automated tools, with a view to highlighting both their outstanding potential as well as their drawbacks (in terms of transparency and non-discrimination). The third work package will build upon the theoretical and technical findings of the first two stages to critically discuss the trends currently developing in the European context with respect to this sensitive matter (think, for instance, of the recent EU Digital Services Act proposal).

Empirical research indicates that - like jurors, lawyers, and people in general - judges frequently rely on intuitive processes to make case decisions. Although this automatic way of thinking can be surprisingly accurate, it can also lead to severe and systematic errors, especially in the courtroom. Evidence shows that judicial decisions are also prone to implicit biases, such as ingroup favoring, anchoring, base rate neglect, hindsight, and racial bias. Given the high stakes in judicial decision-making not only for the parties directly affected but for society in general, the question of improving the quality of these decisions is important. Judicial review, workload reduction, adequate legal education and training are some of the methods commonly pointed out to reach this goal. Judicial decision-making involves multiple tasks, each of which use various cognitive and emotional processes to different degrees. Albeit there are a growing number of investigations exploring the possibility of applying neurotechnologies and psychopharmaceuticals to enhance cognitive function, there is virtually no research exploring the possibility of using these techniques to enhance judicial cognition. Thus, the proposed project aims at filling this gap in scientific literature. It investigates the ways in which traditional and technological cognitive enhancement could improve the quality of judicial decisions. It also analyzes the normative framework for judges to undergo cognitive enhancement, investigating whether there is a moral duty or a legal obligation to do so. As an interdisciplinary project, it will combine ethical and legal theory with elements from cognitive psychology and neuroscience.

My research is about exploring Natural Language Processing methodologies for the detection of argumentative patterns and rules in argumentative and legal texts.
My aim is to explore Natural Language Processing (NLP) methodologies to predict and classify argumentative linguistic patterns (such as Argumentation Schemes) and rules (e.g. deontic rules). Among the aims of my research, there is the goal of extracting textual patterns in the legal domain.
During the last years, I employed both symbolic and sub-symbolic methodologies: vector embeddings, TFIDF, Tree Kernels, Transfer Learning, Neural Networks with this purpose.
Meanwhile, I wrote nearly 10 scientific publications and I have been invited in several international conferences related to the field of Natural Language Processing, Argumentation and Artificial Intelligence.
During my PhD, I worked in international projects such as MIREL and I was a visiting researcher at the Zhejiang University (China).
My PhD is carried on in co-tutelle between the University of Luxembourg and the University of Bologna.