Can AI Help Cut Plastic Waste From the Food System? | Civil Eats

Can AI Help Cut Plastic Waste From the Food System?

U.C. Santa Barbara researcher Nivedita Biyani believes solutions start with the right data. A new tool uses machine learning to estimate how different policy interventions can reduce plastic pollution through 2050. 

A version of this article originally appeared in The Deep Dish, our members-only newsletter. Become a member today and get the next issue directly in your inbox.

In April, representatives of more than 150 countries traveled to Canada to hammer out the details of an international, legally binding treaty to tackle the plastic crisis. It was the fourth of five negotiation sessions, with the process set to be completed later this year.

If all goes according to plan, the result will be a Paris Agreement for plastic.

“Negotiators need to recognize that plastic pollution is an accelerating global crisis that cannot be solved with fragmented national approaches.”

Just like during international climate summits, the companies driving the crisis are showing up in numbers to try to shape the outcome of the negotiations. Also familiar: Many advocates and experts say that treaty progress is not moving fast enough given the urgency the situation demands. In a press release commenting on the April meeting in Ottawa, World Wildlife Fund (WWF) International pointed out that more than 15 million tons of plastic entered the ocean during the week of negotiations.

“Countries have made important progress in Canada with constructive discussions on what the treaty will actually do, but the big decisions still remain: Will we get the strong treaty with common global rules that most of the world is calling for, or will we end up with a voluntary watered-down agreement led by least-common-denominator values?” said Eirik Lindebjerg, WWF International’s global plastics policy lead. “Negotiators need to recognize that plastic pollution is an accelerating global crisis that cannot be solved with fragmented national approaches.”

Nivedita Biyani, an expert on plastic waste, was in Ottawa attempting to provide policymakers with data they could use to make smarter decisions.

Biyani has been working to understand and improve waste management for about 12 years, in poor neighborhoods in India that lack infrastructure, for the government of Singapore, and now as a researcher at U.C. Santa Barbara (UCSB), where she looks at “mass flows” of plastic—aka “how materials travel through production to end-of-life and become waste.”

Her latest project: the Global Plastics AI Policy Tool, developed with Douglas McCauley, the director of UCSB’s Benioff Ocean Science Laboratory, and Roland Geyer, a prominent industrial ecologist.

Nivedita Biyani

Nivedita Biyani

Building on a 2017 paper Geyer published that estimated how much plastic had ever been made, the team set out to show what kind of impact 11 policy interventions—including capping production, taxing plastic packaging, and investing in recycling infrastructure—would have on reducing plastic pollution through 2050. To do so, they used a machine learning model, a form of artificial intelligence, that included GDP and population data to model future years.

The tool estimates that in 2024, 129.7 million metric tons of plastic waste will be incinerated, 173.6 million metric tons will end up in landfills, and 73.5 million metric tons will end up in the environment. From there, without drastic changes to business as usual, the numbers just keep climbing.

Recently, Biyani spoke with Civil Eats to explain how policymakers (and others) might use the tool to reduce plastic waste from every sector, especially the food system.

Can you walk us through how this works?

We start the conversation with mismanaged plastic waste. In this model, mismanaged means it’s in the larger environment, either on land or in the sea, basically not being managed the way it should be.

Click image for a larger version (opens in a new window).

So, what you’re seeing on this axis is you have 2011, and here you have 2050. You can look at, under business as usual, this is the trajectory. Here we are today, at 2024, and we’re at roughly 60-70 million metric tons of mismanaged plastics waste into the larger environment. If we do not do anything, we are likely to reach about 121.5 million tons of mismanaged plastic waste by 2050.

We’ll bring the news to you.

Get the weekly Civil Eats newsletter, delivered to your inbox.

Click image for a larger version (opens in a new window)

(On the left side above,) you have a collection of 11 different policies like, for example, reduced single-use packaging, reduced additives, a cap on virgin polymer production, or implementing a minimum recycled content. You can even toggle the percentage. So, at its highest, for example, implementing a minimum recycled content mandate would reduce the mass of mismanaged plastic waste at 2050 from 120 million metric tons to 64. And we can toggle all the different policies and see what we get. [If we do everything], we get to something like 17 million metric tons, which is as low as it goes.

When you look at this list of policies governments could implement and how each could reduce mismanaged waste, are there any that stand out as being the most effective at reducing the most waste?

Absolutely. Minimum recycled content [a little more than halfway down the list] is one of the biggest ones that actually reduces mismanaged plastic waste. This is saying that every piece of plastic that we put out should have a [20-40%] minimum recycled content of plastic in it. So, this is not only making sure things get recycled, but that the recycled mass is incorporated into new products. This is a big one.

Another one would be to cap virgin plastic production at 2025 levels. If we capped it, we would see a pretty significant delta, about 31 million metric tons of reduction of mismanaged plastic waste.

And then the other one would be waste infrastructure. This is really speaking to the fact that we need to implement systems and incentives to collect whatever we sell. That has a really big effect as well.

What I’m trying to say is that waste is actually a supply chain. No one thinks about it that way. It’s material, it’s useful material, and in this day and age where right now everything is so expensive . . . why are we wasting it?

We cannot keep doing waste management the way we’ve been doing it since the 1950s. Every single industry has changed, has had a disruption. Almost nothing we do right now is like we did in the 1950s except for waste management. Why has that not changed?

When you look at what is being recycled, even after all this time, it’s almost nothing, and most plastic is actually not recyclable. Is it really possible to get to numbers like 40 percent recycled content in all new plastics?

There’s one part of this process, which is the modeling aspect of it, the mass flows. The other is implementing the policies we’ve modeled, and that’s a whole different conversation. When I was working for the government of Singapore, they had a lot of trouble trying to implement some of the policies we’ve talked about. They’re not that easy, and it might be easier for some countries to implement than others.

One criticism I’ve seen of the treaty negotiations so far is that there is a lot of more emphasis on recycling and reuse and not as much on capping production. Based on the data, is it more important to cut production, and can we fix the plastic waste problem without capping? 

Thank you for being a loyal reader.

We rely on you. Become a member today to read unlimited stories.

We could get to 20 million metric tons of mismanaged plastic waste without capping production, but it’s like trying to get all the water out of the house without turning off the tap.

A lot of the plastic waste in the food system is packaging. Is there a specific policy solution that shows the most promise in that sector according to your model?

Packaging reuse on its own does not have a very large impact on mismanaged plastic waste, but packaging reuse with a minimum recycled content mandate would result in a staggering decrease [from 121.5 to 56.5 million metric tons] in mismanaged plastic waste.

This is saying, “OK, how can we extend the life of one plastic packaging to not just one use?” It’s saying, “Maybe we can get eight to 10 uses out of a plastic packaging and then send it for recycling.” And if you can do that, look at the effect you have. It’s quite staggering. It’s more than half in reduction as opposed to only reuse.

For example, if a very well-known coffee company that sells coffee everywhere in America would implement reuse and collect back the coffee cups and then send those for recycling after eight to 10 uses, potentially even 20 depending on the robustness of the plastic, then you would have a much bigger reduction.

This interview has been edited for length and clarity.

Lisa Held is Civil Eats’ senior staff reporter and contributing editor. Since 2015, she has reported on agriculture and the food system with an eye toward sustainability, equality, and health, and her stories have appeared in publications including The Guardian, The Washington Post, and Mother Jones. In the past, she covered health and wellness and was an editor at Well+Good. She is based in Baltimore and has a master's degree from Columbia University's School of Journalism. Read more >

Like the story?
Join the conversation.

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.

More from

Food Policy



Zero-Waste Grocery Stores in Growth Mode as Consumers Seek to Ditch Plastic

Inside a re_ grocery store in the Mar Vista neighborhood of Los Angeles. (Photo courtesy of re_grocery)

Pesticide Industry Could Win Big in Latest Farm Bill Proposal

Restaurants Create a Mound of Plastic Waste. Some Are Working to Fix That.

What Happened to Antibiotic-Free Chicken?

hickens gather around a feeder at a farm on August 9, 2014 in Osage, Iowa. Photo by Scott Olson/Getty Images