ESRs on their first day in Kraków

A Recap of LeADS Training Module 5

/in , /by

From the 14th to 23rd of September 2022, the 15 ESRs met again for their fifth and final training module. This time they had the chance to meet in the beautiful city of Kraków at Jagiellonian University which is one of the seven beneficiaries of the project. The LeADS training program is structured around several training modules that together aim at training a new generation of researchers as legally attentive data scientists that become experts in both law and data science capable of working within and across the two disciplines. Whereas the fourth training module in Crete focussed on the computer scientist perspective, this training module focussed again more on the legal perspective.

Training Week 1: On Data and Ownership

The first week kicked off with issues surrounding propertization of information. Ewa Laskowska-Litak from Jagiellonian University introduced the ESRs to existing legal regimes as well as technical solutions that can provide legal or de facto protection to data. Dariusz Szostek and Rafał Prabucki introduced the ESRs to benefits and risks of blockchain as an opportunity for privacy and intellectual property rights. Furthermore, Adrianna Michałowicz presented and discussed current approaches of two key deliverables from the European Strategy for Data: The Digital Governance Act as well as the Data Act and how they reflect the latest European approach to data governance. Marietjie Botes from Luxemburg University presented legal approaches to ownership and how prior to the European Strategy for Data the Commission reflected on introducing new property rights in machine-generated non-personal data.

ESRs on their first day in Kraków

ESRs on their first day in Kraków

Katarzyna Południak-Gierz from Jagiellonian University focussed her presentation more on the consumer perspective and how personalisation techniques (e. g. behavioural or contextual tailoring) are used on them online. Together with Fryderyk Zoll, the ESRs had the opportunity to discuss the methodology of legal research and difficulties they have encountered throughout their first year of research. A completely different topic was subsequently presented by Katherine Lee, research engineer at Google Brain. In her talk on language models, i. e. models that learn a probability distribution of a sequence given the previous tokens, she elaborated on how they are trained with data as well as how they might pose privacy risk if they leak information. The first training week ended with a session by Marietjie Botes on discrimination and debiasing of algorithms.

Training Week 2: The Next Steps of the LeADS Project and Conference on Human Rights Centred AI

Week 2 kicked off with a session by Arianna Rossi from Luxemburg University on ethical and legal aspects of social media data mining from a researcher’s perspective. In her lecture on transparency and algorithmic decision-making, Agnieszka Jabłonowska focussed on transparency, its promises as well as its shortcomings. Using the example of consumer law, she discussed with the ESRs the efficiency of newly imposed transparency obligations on online platforms with regard to the disclosure of ranking parameters. Looking at the concrete example of a platform where consumers can reserve hotels, the group concluded that in its current implementation the imposed transparency obligations might not achieve their goal of informing consumers on why certain offers are ranked above o thers. An IP law perspective was adopted in the following lectures. Michał Wyrwiński explained approaches to “Protection of Image” in different jurisdictions. Furthermore, Żaneta Zemła-Pacud and Gabriela Lenarczyk discussed approaches to propertization of information in life sciences and to what extent the IP system already provide control over data.

Patricia Thaine adopted again a more technical perspective in her presentation on privacy enhancing technologies such as differential privacy or synthetic data. Mohamed Ali Kandi from Toulouse University introduced the ESRs to the fundamentals of networking, e. g. by presenting to the researchers the technical structure ‘behind’ the internet. Claudia Castillo Arias from INDRA, one the LeADS partners that is closely involved in LeADS activities, provided ESRs with insights into their business activity in her lecture on Military operations planning: difficulties and challenges. On Wednesday 21st, all ESRs attended the conference “Designing Human Rights Attentive AI – An Interdisciplinary Perspective” co-organized by LeADS and LiderLab of Scuola Superiore Sant’Anna. The conference perfectly reflected the interdisciplinary approach of the LeADS project with diverse presentations on topics such as “Can we trust Fair-AI?” (Salvatore Ruggieri – University of Pisa), “Ethical and Legal Issues in Accessing Biobank Data to develop AI tools” (Andrea Parziale – Maastricht University) or “Explaining AI in Cyber Defence” (Marco Antonio Sotelo – Indra).

Conference on Human Rights Attentive AI

Conference on Human Rights Attentive AI

On the day after the conference, ESRs and beneficiaries of the LeADS project met to discuss and evaluate the overall progress of the project. The training in Kraków constituted the last training module for the ESRs. The next time, the ESRs will come together in December in Toulouse for the Technical innovations In Law Laboratories (TILL) workshop, where they will work on challenges that the non-academic partners of the LeADS project are confronted with in their business activity.

 

 

 

LeADS Training – Poster Session and the Transition from Theory to Practice

/in /by

From the 19th to 25th of June 2022 the 15 ESRs met again for their fourth training module. This time they had the chance to meet on the beautiful island of Crete, Greece. Whereas the previous modules in Pisa focussed on the legal perspective, this training focussed more on the computer scientist perspective.

 

LeADS at AIAI: Panel discussion and Poster Session

The training week kicked off with a workshop on “Best Practices for the development of intelligent and trustworthy algorithms and systems” at the 18th International Conference on Artificial Intelligence Applications and Innovations (AIAI). The workshop was divided into two separate panel discussions. The first panel, moderated by Imge Ozcan from VUB, the speakers Katerina Demetzou (Future of Privacy Forum), Paul De Hert (VUB), and Afonso Ferreira (CNRS, Institut de Recherche en Informatique de Toulouse) discussed topics surrounding “Data Ownership, Privacy and Empowerment”. During their contributions the speakers addressed questions such as whether or not a potential data ownership right could integrate into existing data protection law. The second panel, moderated by Giovanni Comandé from SSSA, was entitled “Trustworthy Data Processing Design”. The speakers Jessica Eynard (Toulouse Capitole University), Gabriele Lenzini (University of Luxemburg), and Salvatore Rinzivillo (Italian National Research Council) debated what kind of legal, ethical, and technological framework is needed in order to ensure a trustworthy data economy that increasingly is in demand for secondary data uses.

Furthermore, prior to the conference all ESRs prepared posters to visualize the most important results of their research. All posters were exhibited and accessible to visitors from the AIAI conference which allowed the ESRs to engage and discuss their research with the international audience from the conference.

Training Week: From Theory to Practice

Followed by this engaging introductory sessions, the ESRs presented the results from their individual research they had been conducting throughout the first months from the LeADS project. All 15 ESRs are divided into 4 Crossroads, i. e. major challenges that still need to be addressed in data-driven societies: (1) Privacy vs Intellectual Property (2) Trust in Data Processing & Algorithm Design (3) Data Ownership (4) Empowering Individuals. Each ESR has furthermore been assigned an individual research topic. During the first months of their research journey, the ESRs wrote a State-of-the-Art analysis of their topic which constitutes the scientific foundation for the upcoming research and cross-collaboration between the researchers. During 4 sessions each ESR from each crossroad had 20 minutes to present and discuss their results.

Michael Filippakis from University of Piraeus subsequently discussed how data science impacts society and can support policy-making processes and benefit innovation. In a hands-on session he introduced the ESRs to time series analysis with R programming language. Ilias Maglogiannis from University of Piraeus in his lecture on pervasive health management technology and data analytics gave several practical examples how technology is used to analyze health activities, for instance, in the case of smart clothing to monitor athletic activities. Giovanni Comandé from SSSA talked about (challenges to) secondary uses of data and elucidated on relevant GDPR provisions. Salvatore Rinzivillo in his engaging lecture discussed with the ESRs how to design explainable AI, i. e. how to incorporate intelligibility (how does it work?) and accountability (who is responsible for?) in AI systems.

The last 3 days from the training where dedicated to two practical sessions. First, Mohamed Ali Kandi from UT3 introduced the ESRs to the creation and deployment of smart contracts with solidity. Followed by this introduction the ESRs were divided in groups consisting of both lawyers and computer scientists – reflecting the interdisciplinary approach which the LeADS project epitomizes. These interdisciplinary groups then had to write and deploy a smart contract which allowed a controller to assign and remove the right to data subjects to rectify their data. Second, Itzel Vazquez Sandoval from University of Luxembourg introduced the ESRs to the design and analysis from security protocols. During the subsequent practical session, the ESRs were introduced to the tool ProVerif and had to analyse a bank statement request protocol. In groups the ESRs had to identify potential weaknesses for attacks that could lead to security breaches in the design from the protocol.

Both exercises strengthened the understanding from the ESRs how solutions to GDPR requirements can be designed, implemented, and analysed and thereby putting a dialogue between computer scientists and lawyers into practice.

ESR Presentation during LeADS mid-term meeting

The Legal Perspective in LeADS – Training in Pisa, Italy

/in , /by

From 28th March to 9th April 2022, the 15 ESRs met again for two weeks in Pisa at Scuola Superiore Sant’Anna (SSSA) for their third training module of the LeADS project. Whereas the first training modules in Novembre-Decembre 2021 involved mainly the discussion of subjects related to computer and data science, this training module focussed on the legal perspective.

The cross-interdisciplinary training program for the 15 ESRs constitutes a fundamental part of the LeADS project. Being capable to fully understand the challenges posed by the digital transformation and data economy requires knowledge in different fields such as computer science, law, or economics. The LeADS training program is therefore structured around six modules that together aim at training a new generation of researchers as Legality Attentive Data Scientists that become experts in both law and data science capable of working within and across the two disciplines.

Training Week 1: Digitalisation, the Law and Challenges Related to its Enforcement

The first week began with a research design course on how to assess the societal relevance of research questions taught by Madeline Polmear from Vrije Universiteit Brussel (VUB). The following lectures were marked in particular by introductory courses to EU and international law for non-experts to level the playing field between lawyers and computer scientists in the group. Fryderyk Zoll from Jagiellonian University (JU) introduced the researchers to EU anti-discrimination law and how it translates into legislation that prohibits, for instance, unjustified geo-blocking to prevent discriminations based on customers’ nationality or their place of residence. A more general introduction to international and European protection of human rights was given by Paul De Hert (VUB). In addition, an introduction to human rights theory and how human rights law developed since the creation of the first international instruments, De Hert gave an outlook on possible new fundamental rights that might emerge in the digital era.

Gloria González Fuster (VUB) zoomed into the regulatory framework that surrounds a fundamental right that is of particular importance for the LeADS project: the right to data protection. In her engaging presentation she gave a short overview on the history of the right to data protection and presented both the key players (data subjects, data controllers & processors, data protection authorities) as well as the main data protection principles that lie at the heart of the protection regime of the GDPR. Vagelis Papakonstantinou (VUB) introduced the researchers to another important branch of law in the context of digitalisation: cyber security law. In his lectures he presented EU law and policy with a particular focus on legislative instruments such as the NIS directive or the EU Cybersecurity Act.

Besides various introductory lectures to different fields of law, the researchers had the opportunity to gain practical insights into enforceability of technological regulation in the EU, thanks to the intervention by Anotonio Buttà, chief economist at the Italian competition authority AGCM (Garante della Concorrenza e del Mercato). He elucidated how the competition authority had to set up a diverse team of computer scientists, data scientists, economists, and lawyers in order to address challenges and identify collusive and discriminatory practices in an environment characterized by the implementation of new technologies like machine learning algorithms.

The researchers had furthermore the opportunity to present and discuss the advancement of their work within the foESR Presentation during LeADS mid-term meetingur crossroads of the LeADS project (1. Privacy vs Intellectual Property 2. Trust in Data processing & Algorithmic Design 3. Data Ownership 4. Empowering individuals), i. e. challenges that still need to be addressed in data-driven societies. Since all ESRs are placed in beneficiary institutions in six different countries, the meeting in Pisa constituted a valuable opportunity to discuss and give meaningful feedback in person on each other’s research. The first training week ended with the mid-term meeting where the general advancement of the LeADS project was discussed together with the project officer from the European Commission and where the researchers had the opportunity to present themselves and their individual research.

Training Week 2: The Law and Challenges Posed by New Technologies

During the second training week Giovanni Comandé (SSSA) ensured that the perspective would not only remain interdisciplinary but also cross-cultural, by introducing the researchers to American and Chinese approaches to privacy. The international perspective was maintained in the class by Maria Gagliardi (SSSA) who presented the complex rules of the GDPR that apply to international transfers of personal data.

Furthermore, Giovanni Comandé explained how new computing capabilities and processing technologies transformed us into a networked, data driven, classifying society, where data is increasingly being exploited and used, for instance, to make predictions about individuals’ future actions and to increasingly employ nudging techniques that impact our decision-making behaviour. Questions that will have to be dealt with would therefore concern the frontier between techniques that merely support the decision-making process versus techniques that coerce somebody to take a specific decision. Arianna Rossi and Marietjie Botes from University of Luxembourg (LU) further elaborated on this topic by introducing dark patterns and ethics of online manipulative behaviour.Presentation on Dark Patterns

Gianclaudio Malgieri (VUB) elucidated how big data and AI technologies have created new risks for data subjects since discrimination would be possible based on new “affinity profiling” characteristics. The AI ecosystem would thus lead to layers of vulnerability based on specific contexts.

Caterina Sganga (SSSA) presented in her lecture how AI technology challenges the regulatory framework of intellectual property – would it be possible to conceive machines as an inventor or author and what happens if AI agents infringe IP rights?

Finally, one of the latest legislative proposals, the AI Act, was presented in detail by Giovanni Comandé and enabled discussions on how EU policy attempts to respond to challenges posed by new technologies.

The next training module in June 2022 will take the ESRs to Greece where, in addition to courses in data management and analytics, the ESRs will have the possibility to present and discuss in detail the findings of their research which they have gathered within the first months of their participation in the LeADS project.