Workshop @ Comète on Ethical AI

November 23-24, 2023 - Palaiseau, Paris, France

About

The research team Comète (Inria Polytechnique) is launching a call for proposals for a 2 days workshop on Ethical Artificial Intelligence, that will take place on 23th-24th November 2023. For this second session, we focus on the intrinsic trade-offs that lie at the heart of Ethical AI. As we observe encouraging progress in domains such as Fairness, Privacy, Explainability, and Robustness, we also see tensions between different goals of Ethical AI. The goal of this Workshop is to have a higher-level overview of the challenges and achievements of trying to satisfy multiple desiderata of Ethical AI at once. We seek to bring together researchers from diverse disciplines in computer science, social sciences, law, and industry practitioners. The possible topics include, but are not limited to trade-offs between:

  • Fairness and accuracy
  • Privacy and accuracy
  • Causality in Ethical AI
  • Explainability and accuracy
  • Robustness and accuracy
  • Privacy and fairness
  • Technological progress and following ethical guidelines
  • Technological progress and job loss
  • Cultural values and objectivity
  • Open source and regulation
  • Safety and performance

Join us for interdisciplinary discussions, thought-provoking conversations, knowledge-sharing, and friendly coffee breaks :). Last but not the least, the workshop is getting famous for French gourmet dinner for the participants!

If you have any questions please contact szilvia.lestyan@inria.fr.

Key dates

  • Abstract submission: 09th October 2023 19th October 2023
  • Confirmation of acceptance: 21th October 2023
  • Workshop: 23th-24th November 2023

Keynote Speakers

Avatar

Ferdinando Fioretto

Assistant Professor of Computer Science, University of Virginia

Avatar

Giada Pistilli

Principal Ethicist, Hugging Face.

Avatar

Ljupcho Grozdanovski

Permanent Research Associate, University of Liège

Programme

Thursday
Time Event
10:30 – 12:00 Welcome Brunch
12:00 – 13:15 Keynote: Ljupcho Grozdanovski – University of Liège
The Explanations One Needs for the Explanations One Gives: Thoughts on the Epistemic Link between Explainable AI and Causal (Evidentiary) Explanations under the EU’s AI Liability Regulation
[Abstract]
13:15 – 13:30 Short coffee break
13:30 – 13:50 What is the meaning of “bias” in AI?
Ambre DavatGRESEC / Chaire Ethique & IA (Institut MIAI)
[Abstract]
[Presentation]
13:50 – 14:10 The need for an interdisciplinary approach to AI research Developing a research ethics framework for AI
Anais ResseguierTrilateral Research
[Abstract]
[Presentation]
14:10 – 14:30 What ethics does neuroethics bring to bear on the issue of convergence of AI and neurosciences?
Tabouy LaureUniversite Paris-Saclay, CESP INSERM U1018
[Abstract]
14:30 – 14:50 Dual Use Concerns of Generative AI and Large Language Models
Laurynas Adomaitis, Alexei GrinbaumCEA/Saclay
[Abstract]
[Presentation]
14:50 – 15:15 Break
15:10 – 15:30 On the Impact of Local Differential Privacy on Fairness: A Formal Approach
Karima MakhloufInria and École Polytechnique
[Abstract]
[Presentation]
15:30 – 15:50 Differential Privacy has Bounded Impact on Fairness in Classification
Michaël PerrotInria
[Abstract]
[Presentation]
15:50 – 16:10 Enhancing Metric Privacy With a Shuffler
Andreas AthanasiouPhD Student at Inria/École polytechnique
[Abstract]
[Presentation]
16:10 – 16:30 On the Complexity of Differentially Private Best-Arm Identification with Fixed Confidence
Achraf AzizeInria
[Abstract]
[Presentation]
16:30 – 16:50 Differential privacy surprises: utility is not (always) monotonic on epsilon
Natasha FernandesMacquarie University
[Abstract]
[Presentation]
16:50 – 17:15 Break
17:15 – 18:30 Keynote: Ferdinando Fioretto – University of Virginia
Privacy and Fairness in Societal Systems
[Abstract]
18:30 - 19:00 Free discussions and going to the restaurant
19:00 - ? Dinner at Living-Room (approx 500m walk from INRIA building)
Friday
Time Event
08:30 – 09:00 Welcome Coffee
09:00 – 09:20 A Taxonomy of Security Threats, Vulnerabilities, and Controls of AI Systems
Yusuke KawamotoAIST
[Abstract]
09:20 – 09:40 Legal regulation of AI and morality in the context of natural law and legal positivism
Arturas GrumulaitisVilnius University
[Abstract]
[Presentation]
09:40 – 10:00 When artificial intelligence goes fishing: On the importance of complete and representative training data of AI-driven competition law enforcement
Jerome De CoomanUniversity of Liege
[Abstract]
[Presentation]
10:00 – 10:20 Performativity and Prospective Fairness
Konstantin GeninUniversity of Tübingen
[Abstract]
[Presentation]
10:20 – 10:40 CALIME: Causality-Aware Local Interpretable Model-Agnostic Explanations
Martina CinquiniUniversity of Pisa
[Abstract]
[Presentation]
10:40 – 11:00 Break
11:00 – 11:20 How safe are LLMs compared to search engine autocompletions? A bias check-up
Alina LeidingerUniversity of Amsterdam
[Abstract]
[Presentation]
11:20 – 11:40 When mitigating bias is unfair: predictive multiplicity in algorithmic fairness
Thibault LaugelAXA
[Abstract]
[Presentation]
11:40 – 12:00 Active Fourier Verifier: PAC Estimation of Model Properties with Influence Functions and Fourier Representations
Ayoub AjarraInria Lille - Scool
[Abstract]
[Presentation]
12:00 – 12:20 Leveraging fairness to quantify attribute inference attacks success
Jan AalmoesINSA Lyon
[Abstract]
[Presentation]
12:20 – 12:40 Mitigating inherent biases in language models by reinforcement learning
Miguel CouceiroUniversité de Lorraine, CNRS, LORIA
[Abstract]
[Presentation]
12:40 – 13:45 Lunch
13:45 – 15:15 Keynote: Giada Pistilli – Hugging Face
Conversational AI Ethics: Philosophy and Real-World Case Studies
[Abstract]
15:15 – 15:30 Break
15:30 – 17:00 Round Table
17:00 – 18:00 Free discussions with coffee and snacks

Round table

Invitees

Avatar

Anastasia Stasenko

Research Lead at opsci.ai.

Avatar

Giada Pistilli

Principal Ethicist, Hugging Face.

Avatar

Ljupcho Grozdanovski

Permanent Research Associate, University of Liège

Avatar

Natasha Fernandes

Lecturer in Cyber Security, Macquarie University, Australia

Moderator

Avatar

Mário S. Alvim

Professor of Computer Science at UFMG (Brazil) / Visiting engineer at Inria Saclay (France)

Venue

Inria Saclay
1 Rue Honoré d’Estienne d’Orves
Bâtiment Alan Turing
Campus de l’École Polytechnique
Palaiseau, France

Organizers

Avatar

Szilvia Lestyan

Post-doctoral Researcher at Inria

szilvia.lestyan@inria.fr

Avatar

Ruta Binkyte

PhD Student at Inria, LIX École Polytechnique

ruta.binkyte-sadauskiene@inria.fr

Avatar

Ramon Gonçalves Gonze

PhD Student at Inria, LIX École Polytechnique and Federal University of Minas Gerais

ramon.gonze@inria.fr

Avatar

Catuscia Palamidessi

Director of Research at Inria

catuscia@lix.polytechnique.fr

Avatar

Bilel Benbouzid

Professor at Université Gustave Eiffel

bilel.benbouzid@u-pem.fr