Regulating Facial Recognition Technologies Based on Artificial Intelligence in a Multilevel Context

Biometric facial recognition technologies (FRT) were one of the most controversial issues during negotiations on the European Union’s AI Act that entered into force on 1 August 2024 as Regulation (EU) 2024/1689. The regulation classifies some FRT use cases as prohibited AI practices – with exemptions for law enforcement agencies (LEA), and others as high-risk AI systems. Some of the provisions have already come into force, with others coming into effect in August 2026. Even if the EU opted for a directly binding regulation, the practical impact of this will largely depend upon the way in which member states make use of the margin of manoeuvre that the regulation leaves them to allow the use of FRT to their LEAs in the specific multilevel context. The purpose of this paper is to address the following question: What is the margin of manoeuvre for national implementing legislation on FRT for its public sector? The paper is based on an interdisciplinary legal and political science approach and takes a specific look at the case of the upcoming German AI legislation meant to regulate issues left open by the EU AI Act.