HUDS Rolls Out New Artificial Intelligence System to Combat Food Waste

Harvard has a food waste problem. University administrators think that artificial intelligence can help.

Harvard University Dining Services is piloting a new AI-powered system, Winnow, at Adams House, Currier House, and Annenberg in an effort to reduce its food waste.

The new pre-consumer waste technology is fed the weekly menu items before analyzing food scraps. Winnow pairs a floor scale and a smart camera to weigh food waste in the compost bin, snap a photo of non-donatable food scraps, and precisely identify the menu item.

HUDS’ partnership with Winnow is the University’s latest step to focus on sustainability — a part of their broader institutional commitment to health, cultural, religious, and environmental sensitivities. This has led to similar changes in the dining halls for the Fall 2025 semester, such as the introduction of reusable takeout containers.

“When it came to environmental sensitivity, food waste was an area we knew we would like to mature our own systems and practices for effectiveness and efficiency,” Smith Haneef, HUDS managing director, said.

HUDS was looking to find an easy, accurate solution to the waste generated by the dining hall’s all-you-can-eat buffet service model. After rounds of interviews with four to five organizations equipped with similar software and technologies, Martin Breslin, director for culinary operations, and Crista Martin, director for strategic initiatives and communications, decided to partner with Winnow.

“After interviewing these companies, we picked Winnow on a couple of criteria, and one is that it has to be easy to use for our staff. It also has to produce very good, accurate data” Breslin said.

HUDS has already been able to gather preliminary data about which food to prepare and how much of it to reduce waste in the third week of Winnow’s pilot period.

“When I pull a report from Winnow, almost every day, I can actually see exactly the food coming back in and in real time can ask our culinary team to adjust production records and minimize waste,” Breslin said.

While the dining staff previously had to rely on handwritten, “cumbersome” production records, “Winnow takes all of that off the team and makes it so easy and much more accurate to gather information,” Breslin added.

The AI-powered system, however, is not entirely without flaws.

“Did the vision AI from Winnow truly capture the ingredient or the item, or is it something else, like, is it macaroni and cheese, or is it chicken noodle soup?” Haneef said. “There have been some errors.”

But Haneef added that human correction at the end of every shift offers “another source of enablement of quick feedback loops to correct the AI or the smart camera.”

While Winnow’s predictive analysis is still limited, the team has already been able to use preliminary findings from the three dining hall pilot locations to assess items that are being overproduced.

“The number one actually is rice,” Breslin said. “We saw it across the three locations and that was a very easy fix — like immediate.”

Alongside using Winnow as predictive analytics to confirm what HUDS “intuitively knew but not quantitatively,” the staff have shortened the menu cycles from four weeks to three, streamlining to include student favorites with a high acceptability factor.

“The four week cycle really stretches and adds lots of items to inventory and so on, which can actually add to increasing waste,” Breslin said. “So the three week cycle menu is actually more efficient.”

Haneef has a “high degree of confidence” in Winnow’s future ability to be scaled across the University upon reassessment on December 20, 2025 at the end of the fall semester.

“The hard internal goal that we’ve set is could we achieve 15 percent food waste production at the kitchen production level,” she said. “If that’s true, we’re gonna have a follow up project team meeting, and we want to scale that again.”

— Staff writer Ava H. Rem can be reached at ava.rem@thecrimson.com. Follow her on X @avar3m.