Search results

Is artificial intelligence reliable? Can it make up information without us realizing it? This second session in the “What If AI?” series sheds light on the phenomenon of AI “hallucinations”—errors that can sometimes seem credible and may have significant consequences, particularly in professional settings.Drawing on real-world examples, such as reports containing false references, the speakers will explain why these errors occur, what their technical causes are, and how to detect them. An interactive activity will invite the audience to distinguish between true and false information in AI-generated responses, while exploring the limitations of solutions such as prompt engineering.The goal: to develop a critical perspective on AI responses and gain a better understanding of its limitations. Je m'inscris

AI is often perceived as neutral and objective. But is that really the case? This third session of the “What If AI?” series examines the biases and discrimination that can be embedded in artificial intelligence systems, drawing on striking examples such as biased recruitment and judicial algorithms.Through demonstrations and interactive experiments, the session aims to debunk the myth of technological neutrality and explore ways to design more equitable and inclusive systems. Je m'inscris

What if you confided your doubts, emotions, or sensitive information to artificial intelligence? This first session of the “What If AI?” series explores the uses of AI in personal and sometimes intimate contexts, where tools like chatbots become trusted confidants. Drawing on real-world examples—including the use of ChatGPT for psychological support or to handle confidential data—the speakers will analyze the technical inner workings of these systems as well as the legal issues related to privacy protection.The audience will be invited to discover, step by step, what lies behind an AI-generated response: which actors are involved, how data flows, and what risks this entails. Through a fun, interactive activity, everyone will be able to better understand the implications of their interactions with these tools and reflect on the boundaries that should be set. Je m'inscris

Artificial intelligence creates, imitates, and transforms content on a large scale. But who owns these creations? And what happens to the works used to train these models? This fourth session of the “What If AI?” series addresses the intellectual property issues related to generative AI.Drawing on recent cases involving the use of copyrighted works to train models, the speakers will explain the tensions between technological innovation and respect for copyright. This session will provide insights into the current challenges facing intellectual property law in the face of these technologies and the implications for creators. Je m'inscris