With the UK’s plastic packaging recycling rate still hovering at 51%, according to the government’s 2025 Waste Statistics, the sector faces an urgent need to close the gap between collection and true circularity.
At this week’s Resource and Waste Management Expo in Birmingham, packaging stakeholders are zeroing in on artificial intelligence as a lever to improve the economics and safety of recycling, while advancing toward net-zero objectives.
Amcor’s sustainability and circular economy director, Mark Roberts, points to work underway at the group’s Leamington Spa plant — the country’s largest post-consumer polypropylene (PP) recycling facility — as evidence of how AI can address systemic inefficiencies. The site processes roughly 50,000 tones of household plastic packaging annually, close to 40% of Britain’s recovered PP. Historically, sortation systems have struggled to cope with the variability of kerbside feedstock, which often contains contaminants and packaging in multiple formats. Manual intervention has been necessary to maintain purity, slowing throughput and elevating safety risks.
To overcome those limits, Amcor has incorporated Tomra’s machine-learning separation equipment, designed to identify polymers with far greater precision than conventional optical sensors. By adapting in real time to new packaging shapes or unexpected contaminants, the system allows operators to focus on oversight rather than repetitive sorting. In parallel, Greyparrot’s AI platform tracks and analyzes material flows, generating data on capture rates, contamination trends, and yield losses. That information is fed back to both waste managers and brand owners, supporting continuous improvements in collection schemes and packaging design.
Roberts argues that coupling such tools with established recycling expertise enables the production of high-grade post-consumer resin suitable for contact-sensitive and industrial uses — an outcome difficult to achieve with legacy technologies alone. The CleanStream process, Amcor’s own proprietary platform, further refines PP fractions to meet strict regulatory and customer specifications, creating feedstock that can displace virgin material at scale.
AI’s role is not confined to the sorting line. Analytics derived from monitoring thousands of tonnes of plastics are beginning to inform policy and investment decisions by mapping how materials move through the value chain. Better visibility over flows and leakage points strengthens the case for collection expansion, particularly in regions where capture rates lag national averages. It also supports benchmarking of recycling rates across facilities, which can drive operational discipline and encourage standardization in packaging formats.
Nevertheless, the trajectory is clear. As algorithms learn from increasingly diverse datasets and equipment prices decline, precision sorting is expected to move from pilot applications to industry standard. For operators like Amcor, integrating AI into core recycling assets is part of a broader strategy to embed circularity into product design and resource management.
