Evaluating Chemical Additive Data Management Systems in AI-Driven Plastic Recycling Services for Food-Grade Applications

Sammanfattning

Recycled plastic offers significant potential to reduce waste, but a key barrier to its use in food packaging is the lack of information about the chemicals present in incoming plastics. AI sorting systems can identify plastic types, but cannot detect chemical additives such as flame retardants and stabilizers added during production. Manufacturers rarely disclose this information due to concerns about insecure data transfer and trade secrets. This thesis investigates strategies to improve chemical additive data management for enabling food-grade recycled plastics. Using a qualitative multiple-case study, semi-structured interviews were conducted with Svensk Plastatervinning (Site Zero), Returpack, and Veolia in Sweden. Service Ecosystem Theory guided data analysis. Four main themes emerged: the inability of sorting systems to detect additives, the storage of chemical data in isolated systems, insecure data transfer, and institutional/economic constraints. The principal finding is that the chemical data gap is primarily a governance, not a technical, challenge. Three institutional changes are recommended: mandatory manufacturer disclosure, standardized data protocols, and economic incentives. A proposed framework integrates Laser-Induced Breakdown Spectroscopy for real-time chemical detection and blockchain for secure data transfer, though both require empirical validation before implementation.