AI Dials Up Circularity

Business and IT leaders increasingly understand that the circular economy delivers a transformative way to improve resource utilization and drive sustainability. Yet, as organizations venture down the path of circularity, problems often appear. Aligning design, manufacturing, reuse, repair and recycling can prove daunting. 

Enter artificial intelligence (AI) and machine learning (ML). “AI has the potential to touch every aspect of a circular economy,” observes Paul Miller, vice president and principal analyst at Forrester Research. “While AI alone doesn’t make any particular aspect of circularity possible, it can pull elements together and drive transformation.” 

What does it take to plug AI into the circular economy and put it to work effectively? There is no simple or universal solution. Moving beyond a streamlined linear economy and creating a more dynamic circular economy also isn’t a given, says Ken Webster, founder of the International Society for the Circular Economy and author of “The Circular Economy: A Wealth of Flows”. 

Circular Logic 

One thing is clear: The need for circularity has never been greater. McKinsey & Company reports that every year, $2.6 trillion worth of material in fast-moving consumer goods -- 80% of the material value -- is thrown away and never recovered. This occurs amid a backdrop of trillions of dollars in squandered revenue opportunities and emerging regulations, particularly for companies doing business in the European Union (EU). 

Related:Will Future AI Demands Derail Sustainable Energy Initiatives?

The inherent complexities of the circular economy are formidable. Too often, Webster says, organizations focus on a narrow spectrum of the circular economy that tilts heavily toward recycling along with product, component, and material recovery. These areas are often easier to address and economically viable -- partly because infrastructure already exists to tackle them. Yet these efforts deliver limited gains and often leave bigger opportunities untapped. 

The good news, Miller says, is that circularity doesn’t require a complete retooling of technology and systems. “It isn’t necessary to abandon everything you have been doing for the past 100 years and turn to entirely different technology and systems to promote circularity,” he says. “It’s mostly about taking existing practices and chaining them together more intelligently and effectively.” 

Ideally, circularity examines issues through a broader lens. How can we reduce the use of raw materials? What are some ways to lower energy consumption? How can we optimize packaging and systems to trim recycling and waste? AI can aid in areas as varied as 3D printing, computer vision, robotics, and the use of blockchain. “When you put all these pieces together, and understand interrelationship, you gain a far more complete picture,” Miller says. 

Related:How AI Impacts Sustainability Opportunities and Risks

Using data analytics tools, machine learning models and digital twins, an organization can begin to fuel efficiency in remarkable ways. For example, an EU study, Ecodesign Your Future, found that 80% of the environmental impact of a product is determined at the design stage. “AI can provide insights into physical characteristics, how people use products, and what happens when products reach the end of their lifespan,” Miller explains. 

Consider this: When consumer goods giant Colgate Palmolive wanted to boost recycling rates for its toothpaste tubes, it turned to AI firm Glacier to obtain detailed data about the way people use and dispose of the tubes. Using a robotic system equipped with computer vision, Glacier grabbed images and other data from inside actual recycling facilities and used the data to build a proprietary AI model. 

The data helped Colgate understand whether tubes were making it to the recycling center in the first place, and whether they were winding up in the correct recycling stream at facilities. Colgate, which had already developed a more sustainable tube made entirely from HDPE, used the information to create new marketing materials that helped educate consumers about tube recycling, says Areeb Malik, co-founder and CTO of Glacier. 

Related:Artificial General Intelligence in 2025: Good Luck With That

“The technology allows recycling facilities to process and sort recyclable waste more efficiently and more predictably,” Malik says. At the same time, “AI-generated waste data provides critical insights for recyclers and manufacturers. It's impossible to improve complex systems and develop a circular economy without strong data.” 

Unpacking AI 

The ability to fill information gaps is AI’s superpower, Miller says. IoT sensors, computer vision, ML, digital twins and other digital technologies can determine everything from product lifecycles to what constitutes optimal design and packaging. This data helps organizations identify ways to reduce the use of raw materials, slash waste, and expand refurbishing and recycling programs. The data can also help companies reformulate and reengineer products for cost savings and improved sustainability. 

AI can also help organizations better understand business models -- such as leasing and subscription-based frameworks -- by mashing up historical data and real-time information. These insights can unlock entirely new markets with new customers, notes Autumn Stanish, a director and analyst in the Research and Advisory practice at Gartner. To put these models into motion, AI can untangle complicated reverse logistics, disassembly, and recycling processes. 

Webster points out that mundane tasks such as recycling grading and sorting are merely a starting point for AI. For example, dynamic pricing models that can aid in reverse logistics are a large untapped opportunity, he says. Webster also believes that as the circularity space matures, new opportunities will arise. For example, “AI along with product passports could help identify counterfeits, irregularities, and flag damaged packaging.” 

Yet AI can also lead to unexpected and counterintuitive outcomes. For instance, one large company found a way to significantly reduce copper on smartphone circuit boards, Miller reports. While executives initially cheered the achievement because it cut Scope 3 emissions, they soon realized that leaving a bit more copper on the circuit board would have been preferable. The problem? The volume of metal had dropped below a threshold considered recyclable. So, in the end, AI had solved one problem but created a new one. 

Circularity = Performance Squared 

Achieving a more advanced circular economy requires a multifaceted approach. Business and IT leaders must effectively integrate and innovate with digital technologies, while also developing new AI models that leverage fresh data points. Additionally, organizations need to foster partnerships and address the current gaps in incentives and infrastructure. As Miller points out, even the most advanced AI can’t help if facilities and infrastructure that support circularity don’t exist. 

There’s also a need to reset thinking. “Many companies are accustomed to laying down all the rules,” Miller says. “But to get to success in the circular economy, they must instead focus on partners, incentives, and a more complete understanding of what’s possible. This can include questioning long-held beliefs and changing business models.”