Workshop Schedule

De-biased, diverse, divisive - On ethical perspectives regarding the de-biasing of GenAI and their actionability

AI tech companies cannot seem to get it right. After years of evidence-based criticism of biases in AI, in particular decision models, LLMs and other generative AI, after years of research and toolbox provision for de-biasing, many companies have implemented such safeguards into their services. However, ridicule and protests have recently erupted when users discovered generated images that were “(overly?) diversified” with respect to gender and ethnicity and answers to ethical questions that were “(overly?) balanced” with regard to moral stances. Is this seeming contradiction just a backlash, or does it point to deeper issues? In this talk, I will analyse instances of recent discourse on too little or too much “diversification” of (Gen)AI and relate this to methodological criticism of “de-biasing”. A second aim is to contribute to the broadening and deepening of answers that computer science and engineering can and should give to enhance fairness and justice.

Room: Linke Aula

Algorithmic Fairness Over Time: Advances & Prospects

Miriam Rateike

Can generative AI-based data balancing mitigate unfairness issues in Machine Learning?

Benoît Ronval, Siegfried Nijssen and Ludwig Bothmann

Governing online platforms after the Digital Services Act: an analysis of the Commission decision on initiating proceedings against X

Matteo Fabbri

Algorithmic Fairness in Geo-intelligence Workflows through Causality

Brian Masinde, Caroline Gevaert, Michael Nagenborg, Marc van den Homberg and Jaap Zevenbergen

Threshold Recourse for Dynamic Allocations 

Meirav Segal, Anne-Marie George and Christos Dimitrakakis

Categorizing algorithmic recommendations: a matter of system, agent, patient

Matteo Fabbri and Jesus Salgado

Exploring Fusion Techniques in Multimodal AI-Based Recruitment: Insights from FairCVdb

Swati Swati, Arjun Roy and Eirini Ntoutsi

Social Assessment and Cultural Resistance: The Public Distribution System in Tamil Nadu, India

Sumathi Rajesh

Beyond the AI hype: Balancing Innovation and Social Responsibility

AI can extend human capabilities but requires addressing challenges in education, jobs, and biases. Taking a responsible approach involves understanding AI's nature, design choices, societal role, and ethical considerations. Recent AI developments, including foundational models, transformer models, generative models, and large language models (LLMs), raise questions about whether they are changing the paradigm of AI, and about the responsibility of those that are developing and deploying AI systems. In all these developments, is vital to understand that AI is not an autonomous entity but rather dependent on human responsibility and decision-making. 

In this talk, I will further discuss the need for a responsible approach to AI that emphasize trust, cooperation, and the common good. Taking responsibility involves regulation, governance, and awareness. Ethics and dilemmas are ongoing considerations, but require understanding that trade-offs must be made and that decision processes are always contextual. Taking responsibility requires designing AI systems with values in mind, implementing regulations, governance, monitoring, agreements, and norms. Rather than viewing regulation as a constraint, it should be seen as a stepping stone for innovation, ensuring public acceptance, driving transformation, and promoting business differentiation. Responsible Artificial Intelligence (AI) is not an option but the only possible way to go forward in AI.

Virginia Dignum is Professor of Responsible Artificial Intelligence at Umeå University, Sweden where she leads the AI Policy Lab. She is also senior advisor on AI policy to the Wallenberg Foundations. She has a PHD in Artificial Intelligence from Utrecht University in 2004, is member of the Royal Swedish Academy of Engineering Sciences (IVA), and Fellow of the European Artificial Intelligence Association (EURAI). She is a member of the United Nations Advisory Body on AI, the Global Partnership on AI (GPAI), UNESCO’s expert group on the implementation of AI recommendations, OECD’s Expert group on AI, founder of ALLAI, the Dutch AI Alliance, and co-chair of the WEF’s Global Future Council on AI. She was a member of EU’s High Level Expert Group on Artificial Intelligence and leader of UNICEF's guidance for AI and children. Her new book “The AI Paradox” is planned for publication in late 2024.

Room: Linke Aula

Mapping Policymakers' and Laypeople's Perceptions of genAI and FPT in the Context of the EU AI Act 

Chiara Ullstein, Michel Hohendanner and Jens Grossklags 

Fair Balancing? Evaluating LLM-based Privacy Policy Ethics Assessments 

Vincent Freiberger and Erik Buchmann 

Room: Atrium Maximum

Unfairness in AI Anti-Corruption Tools: Main Drivers and Consequences 

Fernanda Odilla 

Mapping the Potential of Explainable Artificial Intelligence (XAI) for Fairness Along the AI Lifecycle 

Luca Deck, Astrid Schoemäcker, Timo Speith, Jakob Schöffer, Lena Kästner and Niklas Kühl 

Papers from Lightning Rounds 1&2, Papers from In-Depth Session 1

Various Authors

Salon 3Sein, Große Bleiche 60-62, 55116 Mainz
https://maps.app.goo.gl/2KrkFGffTvzDFBgb8

Room: Linke Aula

A Fair Selective Classifier to Put Humans in the Loop

Daphne Lenders, Andrea Pugnana, Roberto Pellungrini, Toon Calders, Fosca Giannotti and Dino Pedreschi

Fairness Beyond Binary Decisions: A Case Study on German Credit

Deborah Dormah Kanubala, Isabel Valera and Kavya Gupta

Challenging "Subgroup Fairness": Towards Intersectional Algorithmic Fairness Based on Personas

Marie Decker, Laila Wegner and Carmen Leicht-Scholten

Deciding the Future of Refugees: Rolling the Dice or Algorithmic Location Assignment?

Clara Strasser Ceballos and Christoph Kern

How to be fair? A discussion and future perspectives

Marco Favier and Toon Calders

Enabling users’ control on recommender systems for short videos: a design proposal for the implementation of the requirements of the Digital Services Act

Matteo Fabbri, Jingyi Jia and Wolfgang Woerndl

Algorithmic Fairness in Clinical Natural Language Processing: Challenges and Opportunities

Daniel Anadria, Anastasia Giachanou, Jacqueline Kernahan, Roel Dobbe and Daniel Oberski

Trust in Fair Algorithms: Pilot Experiment

Mattia Cerrato, Marius Köppel, Kiara Stempel, Alesia Vallenas Coronel and Niklas Witzig

Identifying Gender Stereotypes and Biases in Automated Translation from English to Italian using Similarity Networks

Fatemeh Mohammadi, Marta Anamaria Tamborini, Paolo Ceravolo, Costanza Nardocci and Samira Maghool

Room: Linke Aula

Building Bridges from and Beyond the EU Artificial Intelligence Act. Regulating AI-based Discrimination in the European Scenario

Marilisa D'Amico, Ernesto Damiani, Costanza Nardocci, Paolo Ceravolo, Samira Maghool, Marta Annamaria Tamborini, Paolo Gambatesa and Fatemeh Mohammadi

Abstract:
The panel aims to discuss the implications following the approval of the EU Artificial Intelligence Act also in light of the additional initiatives ongoing in the European and global scenario (e.g. Council of Europe, UNESCO, United Nations) from the perspective of their adequacy to ensure the fairness of algorithms and to tackle discriminations resulting from the massive resort to AI technologies. By bringing together constitutional law and computer science expertise, the discussion has a twofold aim. On the one hand, the panel aims to illustrate the criticisms that underlie the risks of AI systems from a constitutional and human rights perspective, supported by the analysis of computer science; on the other hand, it aims to explore innovative strategies to promote an inclusive use of AI technologies by designers and implementers in order to enhance their positive potential.

Room: Atrium Maximum [30 participants]

Moral Exercises for Human Oversight of AI Systems

Teresa Scantamburlo and Silvia Crafa

Abstract:
The interactive session addresses the challenge of ethical reflection in AI research and practice. Human judgement is crucial in balancing accuracy and fairness trade-offs and overseeing AI system behaviour. To support ethical reflection and judgement we propose the experience of moral exercises: structured activities aimed at engaging AI actors in realistic ethical problems involving the development or use of an AI system. Participants undergo guided exercises involving scenario analysis, individual and group work, highlighting consensus and divergences on the problem at stake. The initiative will promote the development of moral exercises and help refine the methodology for AI ethics education.

Room: Linke Aula

How the Digital Services Act can enable researcher access to data of the largest online platforms and search engines - interactive session with the European Commission 

EC Joint Research Centre

The Digital Services Act (DSA) entered into full force in February, and aims to create a safer and more trustworthy digital space, where the fundamental rights of all users are protected.  As part of the DSA’s transparency obligations, Article 40 of the DSA establishes the obligation of Very

Large Online Platforms (VLOPs) and Very Large Online Search Engines (VLOSEs) to provide researchers with access to data for the purposes of conducting research that

contributes to the detection, understanding and mitigation of systemic risks in the European Union, such as discrimination and the spread of disinformation.    

In particular, Article 40(12) of the DSA obliges providers of VLOPs and VLOSEs to give researchers access to data that is publicly available in their interfaces. In addition, Article 40(4) of the DSA establishes a data access mechanism through which researchers who undergo a vetting procedure can obtain access to non-public data for the study of systemic risks in the European Union.  

In this workshop, the Commission will present the data access mechanism for vetted researchers and participants will have the opportunity to provide feedback on a detailed proposal for its procedural, technical and operational elements, which is currently being prepared in the form of a delegated act. Participants will get the chance to explore how data access could benefit their research, which challenges they foresee and how these may be overcome. Participants will also hear about how the DSA protects researcher access to publicly available data, and will be able to provide feedback on the data access tools and procedures made available by VLOPs and VLOSEs to this end so far. 

Room: Atrium Maximum [30 participants]

Fairness, or not Fairness, That is the Question. Rethinking Virtual Assistants' Responses From an Ethical Perspective 

Giulia Teverini, Joy Ciliani and Alessia Nicoletta Marino

Abstract: The workshop aims to delve into the intricate aspects of ethics applied to human-computer interaction. Participants will engage in practical activities in which they will analyze conversations with virtual assistants, distinguishing between appropriate and inappropriate interactions. In this context, the ethical principles proposed by the European Commission will be used as guidelines to ensure favorable conditions for the development of trustworthy AI-based systems. Immediately after participants will be asked to propose scenario-based solutions to tackle ethical issues which have emerged in the previous stage. At the end of the workshop, each group will show the result to the audience. The collaborative nature of the workshop aims at fostering critical thinking, practical understanding, and collective reasoning.

Schlossbiergarten, Peter-Altmeier-Allee 1, 55116 Mainz 

https://maps.app.goo.gl/bRui4ticSuAM1urh7

"Society-centered AI: An Integrative Perspective on Algorithmic Fairness"

Abstract:  In this talk, I will share my never-ending learning journey on algorithmic fairness. I will give an overview of fairness in algorithmic decision making, reviewing the progress and wrong assumptions made along the way, which have led to new and fascinating research questions. Most of these questions remain open to this day, and become even more challenging in the era of generative AI. Thus, this talk will provide only few answers but many open challenges to motivate the need for a paradigm shift from owner-centered to society-centered AI. With society-centered AI, I aim to bring the values, goals, and needs of all relevant stakeholders into AI development as first-class citizens to ensure that these new technologies are at the service of society.

Room: Linke Aula

Evaluating Research Results on Fairness Issues of AI-based Social Assessment in Asylum Processes and Integration of Refugees – Science meets Practice

AI-FORA

Room: Atrium Maximum [30 participants]

Striving for Equity: Navigating Algorithmic Fairness for AI in the Workplace 

Thea Radüntz, Martin Brenzke and Dominik Köhler 

Room: Linke Aula

Building job seekers’ profiles: can LLM’s level the playing field?

Susana Lavado and Leid Zejnilovic

FAIRRET: A flexible PyTorch library for fairness

Marybeth Defrance, Maarten Buyl and Tijl De Bie

Unveiling the Blindspots: Examining Availability and Usage of Protected Attributes in Fairness Datasets

Jan Simson, Alessandro Fabris and Christoph Kern

Beyond Distributions: A Systematic Review on Relational Algorithmic Justice

Laila Wegner, Marie Decker and Carmen Leicht-Scholten

Pricing Risk: Analysis of Irish Car Insurance Premiums

Adrian Byrne

20% Increase in Fairness for Black Applicants: A Critical Examination of Fairness Measurements Offered by Startups

Corinna Hertweck and Maya Guido

Integrating Fairness in AI Development: Technical Insights from the fAIr by design Framework

Mira Reisinger and Rania Wazir

Algorithmic Collective Action in Recommender Systems: Promoting Songs by Reordering Playlists

Joachim Baumann and Celestine Mendler-Dünner

Fairness in Graph-Theoretical Optimization Problems

Christopher Hojny, Frits Spieksma and Sten Wessel

It is not about Bias but Discrimination

Chaewon Yun, Claudia Wagner and Jan-Christoph Heilinger

Exploration of potential new benchmark for fairness evaluation in Europe

Magali Legast, Lisa Koutsoviti Koumeri, Yasaman Yousefi and Axel Legay

Room: Linke Aula

Measurement Modeling of Predictors and Outcomes in Algorithmic Fairness 

Elisabeth Kraus and Christoph Kern 

Proxy Fairness under the GDPR and the AI Act: A Perspective of Sensitivity and Necessity 

Ioanna Papageorgiou 

Room: Atrium Maximum

When the Ideal Does Not Compute: Nonideal Theory and Fairness in Machine Learning 

Otto Sahlgren

Inherent Limitations of AI Fairness 

Maarten Buyl and Tijl De Bie