Programme: 

​

09:00–09:30 Introduction of the workshop.
09:30–10:30 Keynote Talk - Laurence Devillers , 30min + 30min Q&A and Discussion.
10:30–11:00 Coffee Break and Open discussion.
11:00–12:30 Talk Session 1 and  Group Activity
12:30–13:30 Lunch Break
13:30–15:00 Talk Session 2 and Group Activity
15:00–15:30 Coffee Break
15:30–16:30 Talk Session 3
16:30–17:00 Workshop Summary and Wrap

​

Keynote:

​

Title - Ethical issues of Generative AI

Laurence Devillers

Abstract: In this keynote, I offer studies and reflections on the ethical issues of generative Artificial Intelligence. The special feature of generative artificial intelligence systems is that they are based on generative models that can produce multiple outputs: generation of text or images for various purposes such as translation, production of computer code, chatbots, decision support and so on. These models, pre-trained on large datasets, can be optimised to produce a new application using little additional data specific to that task. The social and economic impact of generative AI systems is likely to be major in many potential uses, for example in the environment or in healthcare. However, these generative AI systems raise many ethical, epistemological, anthropological, psychological, economic, social, political and cultural questions. Some of these issues will continue to occur as these technologies are put to new uses, and it is not yet possible to predict all the effects they will have on individuals and 
society. Since the end of 2022, economic and political actors in several countries are discussing the impact of language models built with these generative AI systems. Some of these models have an impressive number of parameters. The race for the largest model is ongoing but it is not certain that larger models would deliver higher performance. I was involved as a co-writer of the opinion n°7 of the ethical issues of generative Artificial Intelligence in the CNPEN (National Comity Pilot on Ethics of Numeric). In this opinion, CNPEN focuses on the most 
important ethical issues in the light of current experience with generative AI systems, mainly on language models.

BIO: Laurence Devillers is a Full Professor of Artificial Intelligence (AI) at Sorbonne University and heads the research team on "Affective and social dimensions in Spoken interactions with (ro)bots: technological and ethical issues" at CNRS-LISN. Since 2020, she also leads the interdisciplinary Chair (including economists, linguists and computer scientists) on AI and digital nudge HUMAAINE: HUman-MAchine Affective INteraction & Ethics at CNRS. Her topics of research are Human-Machine co-evolution: from the modeling of emotions and human-robot dialogue to the ethical impacts for society and the risks and benefits of AI. She is involved in the tripartite project LeCycl. She is a member of the National Comity Pilot on Ethics of Numeric (CNPEN). She is president of the Foundation Blaise Pascal on cultural mediation on Mathematics and Computer Science. She is responsible of the JTC21/CEN_CENELEC WG4 on Foundational and Societal Impact of AI that include AI-enhanced Nudging, Trustworthiness AI and "Green" AI. She wrote large-audience books : 'Les robots émotionnels' (Ed. L'Obs., 2020) 
and ´Des Robots et des Hommes: mythes, fantasmes et réalité' (Ed. Plon, 2017).