Algorithmes contre algorithmes : de l’erreur humaine au dysfonctionnement de systèmes automatisés
Fabien Liégeois — 15 janvier 2025
The banking and finance industry was ahead of the curve in reducing human intervention in its operations. As early as the 1970s, communication technologies prepared for the rise of algorithmic trading. Back to the future, machine learning systems differ from rule-based (or symbolic) systems where logic dominates. They rely on data, analyse trends and inferences to generate plausible, but not exact, results. From a legal perspective, the shift from deterministic systems to systems making decisions on a fully or partially autonomous basis raises new interpretative questions. Since our legal system continues to hold users accountable for machine acts or omissions, this contribution emphasizes the leverage effect triggered by the combination of digitization and automation in the case of a mistake. To support our line of reasoning, we will identify three scenarios : contracts concluded between individuals in person, contracts facilitated through a simple digital intermediary, and contracts entered into entirely without human intervention.
Consulter