Circular Data: A Hybrid Intelligence Framework for Data-Driven Disassembly and Repair in the Circular Economy

The transition to a circular economy in electronics is limited by missing product knowledge, low automation, and the absence of trusted mechanisms for secure data sharing. Current approaches such as Digital Product Passports or industrial data spaces improve traceability but do not enable task-level data access suitable for disassembly. This paper presents the Circular Data Framework (CDF), integrating three previously isolated domains: (i) data trusteeship for selective and role-based product data disclosure, (ii) hybrid intelligence using retrieval-augmented large language models and computer vision for generating context-aware instructions, and (iii) hybrid human–AI–robot workcells for ergonomically optimized circular operations. The CDF establishes an end-to-end reference model that links product data governance with AI-driven guidance and robotic execution. We derive a structured research agenda and demonstrate how these components can inform disassembly workflows and ecosystem collaboration. The CDF provides a foundation for secure, scalable, and human-centered circular data infrastructures and addresses the missing link between data, AI intelligence, and robotic execution in circular systems.