ORBIS Project at AI UK Fringe Event
ORBIS, along with other projects funded by the European Commission (AI4Gov, KT4D, Ithaca), is part of a Task Force named AI, Big Data, and Democracy. This Task Force operates in-between interdisciplinary domains, leveraging diverse theoretical perspectives to tackle the complex interplay between AI, big data, and democracy. Informed by democratic theory, it considers principles of participation, representation, and deliberation to comprehend and move across the impacts of emergent technologies on democratic governance.
The Task Force was selected to participate in the AI UK Fringe (4-29 March 2024), an event that brings together leaders in academia from across the UK’s AI ecosystem to showcase, exhibit, and provide updates on their groundbreaking work, with an exploratory, in-depth workshop.
Held on March 7th, this workshop delved into the integration of AI within our societies, mindful of its potential to foster anti-democratic environments for citizens. It tackled a spectrum of ethical and socio-technical challenges to ensure AI’s development, deployment, and usage resonate with human values and foster beneficial societal outcomes, embodying the essence of responsible AI.
Moderated by Grazia Concilio (Associate Professor at Politecnico di Milano, and ORBIS co-coordinator), the first part of the workshop laid the theoretical groundwork for implementing responsible AI in democratic processes through a series of brief keynotes. The session covered the intersection of AI and democracy, diving into risks and opportunities, it looked into Online Deliberation and how it is addressing the gap of bridging theory and practice, it examined the topic of trustworthy and explainable AI in policy making, seeking to enhance AI’s trustworthiness, fairness, and explainability, and finally, it explored practices for supporting a participatory democracy that is inclusive of vulnerable groups. These presentations then fueled a panel discussion that dove into intersectionalities, consistent gaps, and challenges that require attention.
The second part, moderated by Lucas Anastasiou (Research Associate, The Open University) showcased a series of live demonstrations of modern tools for employing AI in deliberative settings, offering attendees a guided experience with these tools.
The workshop was organised by:
- Anna De Liddo, Professor of Human-Computer Interaction and Director of Research, Knowledge Media Institute, The Open University
- Lucas Anastasiou, Research Associate, Knowledge Media Institute, The Open University
- Ilaria Mariani, Researcher, Department of Design, Politecnico di Milano
Agenda:
Part 1
A theoretical grounding on conducting responsible AI in democratic processes, via a series of short keynotes
- AI and Democracy: Risks and Opportunities
Prof Keith Hyams, University of Warwick, UK - Online Deliberation: Bridging Theory and Practice
Prof Anna De Liddo, Open University, UK - Trustworthy and Explainable AI in Policy Making: How can we enhance trustworthiness, fairness and explainability of the modern AI by enabling humans to reason about the outcomes of AI-based models.
Fabiana Fournier, IBM Research, Israel - Aiming for participatory democracy and the inclusion of vulnerable groups: challenges, aspirations and lessons
Michael Bedek and Maria Zangl, University of Graz
The recordings of Part1 is available on Youtube
And here the presentation slides.
Part 2
Demonstration and hands-on experience of a series of technological tools that address this challenge
- BCause: Reasoning for change: A Structured and Decentralised Discussion System for Distributed Decision Making
Lucas Anastasiou, Open University, UK - DemocraticReflection: collective intelligence platform to harness real-time audience feedback over live debates
Anna De Liddo, Open University, UK - DISPUTool: A tool for Mining and Exploring arguments in US Presidential Election Debates, adapted to deliberative democracy debates
Cristian Adriàn Cardellino, Université Côte d’Azur, FR - AI4Gov: A Self-Explained Visualisation Workbench: a Bias Detection Toolkit that empowers policymakers, researchers, and practitioners with understanding tailored for detecting biases in AI systems
Andreas Karabetian, University of Pires, GR - ORBIS: On the development of a dashboard for deliberation analytics
Nikos Alimpertis, Nikos Karacapilidis, Ioannis Livieris, Novelcore, GR
The recordings of Part 2 is available on YouTube
And here the presentation slides.