Separating the privileged wheat from the chaff with AI: A Socio-Legal Exploration of AI-Driven Filtering Systems for Protecting Professional Privilege in Criminal Investigations

This project explores the obstacles standing in the way of protecting digital privileged exchanges in criminal investigations through AI filtering. A high-profile fraud case involving thousands of privileged emails exposed the limitations of traditional keyword-based filtering. Despite AI's potential to enhance efficiency and accuracy, its deployment remains unrealised. Employing empirical socio-legal methods, including interviews and focus groups with police, prosecutors and attorneys, this research will identify legal, technical, and organisational requirements of an effective AI filtering system. This pioneering study will inform the ongoing modernisation of the Dutch Code of Criminal Procedure and safeguard professional privilege in the digital era.