Waste-management facilities are using AI to turn trash into recycled treasure

• Millions of recyclable materials are put into incorrect waste streams every year.
• The UK startup Greyparrot uses AI to help mitigate environmental harm from improper waste disposal.
• This article is part of "Build IT," a series about digital tech and innovation trends that are disrupting industries.

The proverbial "reduce, reuse, and recycle" approach to waste disposal is easier said than done. In the past few decades, global conservation initiatives have tried to promote an eco-friendlier attitude toward recycling, but their impact has been relatively lackluster. Valuable recyclable materials are still thrown in with the trash and sent to landfills or incinerated.

Advanced waste-management technology is an underexamined market that could be the key to stopping precious materials from falling through the cracks. Thanks to the startup boom, it's starting to get more attention.

"If you want to actually have a circular economy, you're going to have to get into mining our waste stream for resources, and as technology develops, that's going to get more and more efficient and more effective," said Steven Cohen, the director of the Earth Institute's research program on sustainability policy and management at Columbia University.

The rate of recycling in the US has grown nearly fivefold in the past 60 years, standing at 32%. The Environmental Protection Agency is aiming to raise that number to 50% by 2030.

Using AI and imaging to sift through trash

Founded in 2019, the UK startup Greyparrot uses specialized tech to enhance waste management. The company's software, installed at waste-management sorting facilities, uses artificial intelligence to identify misplaced materials in waste streams.

Mikela Druckman, a cofounder of Greyparrot, told Business Insider she saw an opportunity to enter waste management after observing a lack of data collection in the industry. She said only a small percentage of processed waste is measured and qualitatively examined. "This led us to see the opportunity of basically digitizing and building the leading waste-intelligence platform," she added.

Greyparrot devices are installed above incoming waste streams and designed to use AI-powered imaging to identify recyclable materials including plastics, aluminum, and paper. These objects can then be sorted out of the waste stream and sent to recycling facilities.

Greyparrot devices also catalog what enters the facility so that waste-facility managers can better understand what they're collecting.

Waste that enters a facility essentially has a "negative value" until it's sorted and sold to recycling facilities, Druckman said. "The role of the waste managers is basically to transform that into a clean, pure commodity," she added.

Greyparrot has 50 devices in 14 countries, including the United Kingdom, South Korea, and the United States. The company sells its devices directly to major waste-management companies.

The company decided against developing hardware like robotics to sort trash. Druckman said she wanted the company's AI tech, analyzers, and software to fit into existing systems.

Data is transforming the way waste-management facilities operate

Since 2020, the UK waste-management company Grundon has been using Greyparrot devices in three of its facilities.

"My colleagues would be looking at ways of how we can improve data collection for what our facilities are doing," said Owen George, the commercial and resource strategy manager at Grundon.

Before installing the Greyparrot system, Grundon's only way of estimating the type and number of materials entering waste streams was to take small samples of collections throughout the facilities and manually go through the materials.

Greyparrot devices gave Grundon's waste-management plants a data visibility they didn't have before. "It even gives us the value to say that we have seen X amount of paper which is valued at X," George said. "We can see how the plants are performing from a revenue perspective."

Grundon's feedback also led to updates in the Greyparrot device's capabilities. When they were installed, the devices were focused on counting materials. "It kind of didn't speak the waste-industry language, which is in weight," George said. A feature that included weight data was later installed.

George described the adjustment to Greyparrot's system as difficult for plant managers at first: The waste-management industry's reliance on manual systems and its limited data collection meant they had to interpret an immense amount of new data. But after they learned the system, their responses changed. "We're at the point now where our people are saying, 'I want to have these units in, because it just makes my life so much easier,'" George said.

AI and robotics could be a new future for waste management

Waste-management facilities are a crucial part of mitigating improper trash disposal, but they're not very attractive to communities. "Nobody wants it located near where they live," Cohen said.

Grundon has tried to demystify the waste-management process by hosting tours of its facilities for locals.

AI development is also promising a brighter future for waste-management facilities and their perception. Druckman said her goal is to create "smart material-recovery facilities" that are constantly "adjusting and in real time adapting to different types of materials and different types of mixes of composition of waste."

Greyparrot recently struck a strategic partnership deal with Bollegraaf, one of the largest waste-management companies, to build facilities designed to be fully integrated with AI systems.

Cohen said a fully automated system of waste collection and disposal — with self-driving vehicles that collect garbage and facilities that operate autonomously — could be made possible through AI and robotics.

He said these could also be used to separate all waste in a single stream at a facility, removing the need for people to sort their own garbage. This, he said, "is probably going to be the most promising technology."

At Grundon, George is already seeing a change in the kinds of jobs becoming available because of AI integration and data expansion. "It's less people with hammers and more people with laptops," George said.

Cohen said that to make progress, the waste-management industry needs people outside facilities to rethink garbage. "That would be the long-term vision: trying to have garbage reconceptualized as a resource as opposed to something that smells and is ugly," he said.