6G and the EU AI Act: Compliance Framework for Network AI Systems

The EU AI Act classifies AI systems by risk level and imposes specific obligations on high-risk systems. This paper analyzes which 6G network AI applications fall under high-risk classification and develops a compliance framework. We find that AI systems managing network slicing for critical services, autonomous network security, and AI-driven resource allocation for emergency services are classified as high-risk. Our framework maps Act requirements to specific technical implementations, enabling operators to achieve compliance while maintaining AI innovation.