In 2023, Michaela Barnet asked study participants to rank the words "reduce," "reuse," and "recycle" in order of environmental impact. More than 78% of participants did it wrong.
The phrase is ordered the way it is for a reason: Reduce first and recycle as a last resort. But decades of recycling campaigns have inverted the hierarchy in public understanding.
"Producers create products that eventually become waste and saddle individuals with the responsibility of disposal," Barnett, an environmental researcher at the University of Virginia, said in releasing her 2023 paper.
Why the hierarchy exists
The Environmental Protection Agency's waste management hierarchy ranks source reduction first, reuse second, and recycling third. Preventing a product from being manufactured eliminates its extraction, production, transportation, and disposal emissions. Recycling it recaptures only a fraction of those costs. The gap between the two strategies is large, and EPA research hasn't changed that conclusion in decades.
What has changed is public understanding. Decades of recycling campaigns — backed by industry groups with a financial interest in continued production — have flipped the order in the minds of most consumers. Barnett's 2023 paper in Nature Sustainability put data behind what environmental scientists had long suspected: Americans treat recycling as the primary solution to a problem it was never designed to solve.
A study by Jesse R. Catlin and Yitong Wang in the Journal of Consumer Psychology found that access to recycling increased resource consumption compared with situations where no recycling option was available. In one experiment, participants used more paper when they could recycle the scraps. In a field test, per-person paper towel use in a restroom rose after a recycling bin was introduced. A 2017 study in the Journal of Marketing Research by Monic Sun and Remi Trudel confirmed the pattern: the positive emotions tied to recycling can overpower the discomfort of waste, and people consume more when a recycling option exists.
The permission structure
Behavioral scientists call this moral licensing. A previous virtuous act activates a positive self-image, which makes a less virtuous one easier to justify. The research shows the guilt associated with waste does not disappear. It gets neutralized by the good feeling of recycling before it can change behavior.
Callie Babbitt, a sustainability researcher at the Rochester Institute of Technology, frames the problem in terms of design. "We're not necessarily designing products at the front end that are well suited for being processed at the back end," she recently told Quartz. The system asks consumers and sorting facilities to manage what manufacturers built without them in mind.
The economics reinforce that dynamic. Virgin plastic costs less than recycled plastic in the United States, a price difference driven by oversupply and falling production costs for new material, according to the Institute for Energy Economics and Financial Analysis. The OECD's Global Plastics Outlook found that only 9% of plastic waste is recycled globally, while 40% of what gets collected for recycling is ultimately discarded as residue. Trevor Zink of Loyola Marymount University and Roland Geyer of UC Santa Barbara demonstrated mathematically in the Journal of Industrial Ecology that without displacing primary production, recycling delays but does not prevent material from reaching final disposal. The climate benefits the industry cites depend on whether recycled material actually replaces virgin output.
That replacement isn't guaranteed.
Where the money is going instead
The recycling industry's answer to these compounding failures is better sorting. Waste Management $WM is spending more than $1.4 billion automating its facilities. Companies deploying computer vision and robotic pickers argue they can push the current rate — only 21% of residential recyclables are captured, according to The Recycling Partnership — substantially higher. The logic is that if consumers won't sort, machines will do it for them.
AMP, a Colorado-based company building AI-powered sorting facilities, has acknowledged what that investment cannot fix. Its systems can identify roughly 90% of materials in the waste stream. But buyers exist for only 50 to 60% of it. Sorted material without a market is still waste.
None of this means the technology is useless. Extracting aluminum, copper, and palladium from electronics waste has real value. Routing textiles toward chemical separation is better than landfilling them. Babbitt sees genuine potential for AI further upstream. Think of smart bins that show commercial kitchens what they're discarding, software that nudges consumers toward repair, or apps that route people to refurbished electronics.
The problem is that none of those applications are where the billions are going. The money is at the back end of the waste stream, sorting material that shouldn't have been produced in the forms it takes. Barnett is direct about why. Without policy that shifts disposal costs back onto manufacturers, she argues, improvements at the sorting stage function less as environmental progress and more as infrastructure that subsidizes continued overproduction.
The EPA hierarchy hasn't changed. Source reduction is still the most preferred strategy, and recycling is still the fallback. The question raised by billions of dollars in new sorting technology is whether the industry is building a more efficient plan or a more comfortable reason not to do the harder thing.