The Center for Data Innovation spoke with Alisa Pritchard, Head of Marketing for Greyparrot. Greyparrot is an AI waste analytics platform that focuses on digitising the waste management industry and tackling the growing waste crisis. Greyparrot equips businesses and regulators with the insights they can use to bolster recycling rates and introduce waste accountability.
Kir Nuthi: How does Greyparrot use computer vision and AI to unlock the financial value of waste?
Alisa Pritchard: Computer vision is an area in artificial intelligence (AI) that uses camera systems to capture and analyse images and videos to identify and categorise objects. When applied to MRFs (Material Recovery Facilities), recent advancements in this technology allow machines to “see” and classify waste at human-level recognition or better. Currently, less than 1 percent of waste is measured by hand. Manual sampling of waste is limited by time and resource constraints and as a result covers a very small proportion of all processed waste. Continuous monitoring of waste through computer vision ensures a much better depiction of true waste composition.
Starting at the recycling plant, accurate, real-time composition data allows waste operators to improve plant efficiency, automate manual sampling and verify the quality of the sorted material in real-time to take corrective action. The data allows waste operators to unlock the true value of their waste and improve efficiency and commercial outcomes.
Furthermore, recycling plants can give objective feedback to those at the start of the chain, whether it be policymakers, local authorities, or packaging designers. Taken together, AI computer vision provides previously unavailable data insights to move the industry from guesswork to a data-driven approach. The industry can have a better understanding of the value of materials flowing through the system and how best to capture that value.
Nuthi: Why does Greyparrot focus on the waste management industry and the regulations surrounding recycling for its data-based solutions?
Pritchard: As it stands, there’s very little information available on the composition of waste. What can’t be measured can’t be managed. These inefficiencies have devastating consequences, with over $82 billion worth of recyclables sent to landfill each year. We believe it is our responsibility to help solve this problem.
What we do is closely tied to regulation. DEFRA (Department for Environment, Food and Rural Affairs) has indicated that “the lack of recent, robust data on the composition of mixed municipal waste streams, both at point of production and at point of disposal, hampers design, implementation, and evaluation of policy.” Furthermore understanding waste composition “is fundamental to the Strategy’s objectives of eliminating avoidable plastic waste over the lifetime of the 25 Year Environment Plan, and eliminating avoidable waste by 2050.”
A key part of the government’s waste strategy is Extended Producer Responsibility (EPR). The data we provide can inform policy decisions on materials and packaging requirements, getting very specific in the ask to packaging producers and ultimately having the data to successfully measure the impact of the EPR scheme to guide further legislation.
Furthermore, regulation places a huge demand and opportunity for the recycling industry. For instance, on April 2022 the UK plastics tax came into effect. Producers are open to fines of £200 per tonne on plastic packaging with less than 30 percent recycled plastic. The aim of the tax is to incentivise producers to use recycled material in their packaging. This unlocks an urgent need for the industry to invest in innovative solutions to increase production of high-quality recycled plastics.
Nuthi: What waste types do you currently analyse through deep learning and what are some other potential applications of your AI in either the waste management or sustainability space?
Pritchard: Over the past two years we have tracked more than ten billion packaging items in waste facilities which allowed us to build highly accurate recognition models across 50+ waste categories including plastics, fibre, and metals.
With this round of funding, we will keep expanding our waste taxonomy to 200+ categories to cover more waste streams including some of the world’s heaviest polluting industries such as construction and demolition.
We will continue to provide our AI Waste Recognition System to waste management companies who operate MRFs (Material Recovery Facilities) and PRFs (Plastics Recovery Facilities). We also work with suppliers of sorting machinery and robotics by providing a AI vision integration product (via our APIs) to improve sorting efficiency and recovery rates.
Nuthi: What are some primary concerns you have encountered regarding the use of AI and deep learning for waste management?
Pritchard: As with many AI systems, there is a risk that the product may automate jobs. But three factors counterbalance the risk.
Firstly, there is a stark labour shortage in this sector.
Secondly, in a traditional industry, automation has proven to heighten productivity across markets; not doing the same in the UK could alternatively risk losing jobs to international firms.
Finally, sorting waste by hand is fraught with a myriad of health and safety risks due to the hazardous materials that end up in the recycling stream.
Nuthi: Now that Greyparrot has closed on £8.9 million in series A funding, what’s the next phase of research and development?
Pritchard: As we’ve mentioned, we will expand our waste taxonomy to 200+ categories to cover more waste streams including some of the world’s heaviest polluting industries such as construction and demolition. We will also expand our product offering to producers and regulators to guide policies on packaging and enhance the overall accountability in the waste value chain.
Last but not least, with this funding round we will be recruiting the best talent across commercial, technical and product roles to prepare for the deployment of hundreds of Greyparrot systems in the coming year. If you are interested, visit our careers page or follow us on Twitter and LinkedIn!