Ensuring the safety of artificial intelligence: a comprehensive analysis and framework

This doctoral thesis aims to provide a thorough examination of the critical issue of ensuring the safety of Artificial Intelligence (AI). As these systems become increasingly integrated into various aspects of society, addressing safety concerns has become paramount. The research explores the potential risks associated with AI, including ethical considerations, unintended consequences, and the challenge of designing systems that align with human values. The study also investigates existing frameworks and methodologies for AI safety and proposes a comprehensive framework that encompasses both technical and ethical dimensions including the automotive example of pedestrian detection systems.