AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
The Risks and Downsides of Suspicion Machines
The chapter explores the concept of 'suspicion machines' in the welfare context, which are machine learning models used to assign risk scores and determine alleged welfare fraud. Concrete examples are given, including a case in the Netherlands where 30,000 families were wrongly accused. The chapter also discusses the challenges in determining the existence and location of deployed machine learning models and highlights the harm and issues caused by AI systems.