Report on Harvard University’s AI Initiative for Food Waste Reduction in Alignment with Sustainable Development Goals

Introduction

Harvard University Dining Services (HUDS) has initiated a pilot program utilizing artificial intelligence (AI) to address significant levels of pre-consumer food waste. This initiative represents a core component of the university’s institutional commitment to sustainability and responsible environmental stewardship. By leveraging innovative technology, Harvard aims to enhance the efficiency of its dining operations, directly contributing to several key United Nations Sustainable Development Goals (SDGs).

Alignment with Sustainable Development Goals (SDGs)

The project is a strategic effort to contribute to the global 2030 Agenda for Sustainable Development. Its primary contributions are focused on the following goals:

• SDG 12: Responsible Consumption and Production: The initiative’s central objective is to substantially reduce food waste through prevention, directly addressing Target 12.3, which calls for halving per capita global food waste at the retail and consumer levels by 2030. The data-driven approach allows for more responsible production patterns within the university’s food system.
• SDG 9: Industry, Innovation, and Infrastructure: By adopting the cutting-edge Winnow AI system, HUDS demonstrates the application of advanced technology and innovation to solve critical sustainability challenges. This fosters a culture of sustainable innovation within institutional infrastructure.
• SDG 2: Zero Hunger: While focused on pre-consumer waste, optimizing food production and minimizing losses are fundamental to creating more efficient and sustainable food systems. This contributes to the broader goal of ensuring food security and ending hunger.
• SDG 13: Climate Action: Reducing the volume of food sent to compost or landfill directly mitigates the generation of methane, a potent greenhouse gas. This action represents a tangible contribution to climate change mitigation efforts.

Technological Implementation: The Winnow System

HUDS has partnered with Winnow to deploy its AI-powered waste tracking system in a pilot program across Adams House, Currier House, and Annenberg Hall. The technology was selected based on its ease of use for culinary staff and its capacity to produce highly accurate and actionable data.

System Operation

1. A smart camera and a floor scale are paired to analyze non-donatable food scraps discarded in compost bins.
2. The system weighs the food waste and captures a photo for analysis.
3. The AI vision technology identifies the specific menu item, referencing a pre-loaded weekly menu.
4. Real-time reports are generated, allowing the culinary team to immediately adjust production records and minimize further waste.

Pilot Program and Preliminary Findings

Initial data gathered during the pilot’s third week has already yielded significant insights, enabling HUDS to transition from intuitive assumptions to quantitative, data-driven decision-making in pursuit of SDG 12.

Key Outcomes

• Data-Driven Production Adjustments: The system immediately identified a consistent overproduction of rice across all three pilot locations. This insight allowed for a simple and immediate correction to production levels.
• Menu Cycle Optimization: To further reduce waste linked to extensive inventory, HUDS has streamlined its menu cycle from four weeks to three. This change focuses on student-preferred items, increasing efficiency and reducing the potential for ingredient spoilage.
• Improved Record-Keeping: The automated Winnow system replaces a cumbersome and less accurate process of handwritten production records, improving data integrity and operational efficiency.

Challenges and Mitigation Strategies

The primary operational challenge involves occasional misidentification of food items by the AI vision system. However, a mitigation strategy is in place.

• Human-in-the-Loop Correction: At the end of every shift, staff review the AI’s classifications and make necessary corrections. This process serves as a rapid feedback loop, enabling the AI to learn and improve its accuracy over time.

Future Outlook and Scalability

HUDS leadership has expressed a high degree of confidence in the program’s potential for a university-wide impact on its sustainability goals.

• Performance Assessment: A formal reassessment of the pilot program is scheduled for December 20, 2025, at the conclusion of the fall semester.
• Waste Reduction Target: The internal goal for the pilot is to achieve a 15 percent reduction in food waste at the kitchen production level.
• University-Wide Expansion: If the pilot successfully meets its reduction target, HUDS intends to scale the technology across all university dining facilities. This expansion would significantly amplify Harvard’s contribution to SDG 12 and its broader commitment to environmental sustainability.

1. Which SDGs are addressed or connected to the issues highlighted in the article?

SDG 12: Responsible Consumption and Production

• The entire article focuses on Harvard University’s effort to reduce food waste, which is a central theme of SDG 12. The initiative described, using the AI-powered system Winnow, is a direct action to create more sustainable consumption and production patterns within the university’s dining services. The text states, “Harvard has a food waste problem,” and the partnership with Winnow is part of the university’s “latest step to focus on sustainability.”

SDG 9: Industry, Innovation, and Infrastructure

• The article highlights the use of innovative technology to solve a sustainability challenge. Harvard is piloting a “new AI-powered system” that “pairs a floor scale and a smart camera to weigh food waste.” This adoption of advanced technology (AI) to upgrade the infrastructure of its dining services and make operations more efficient and sustainable directly relates to fostering innovation for sustainability.

SDG 11: Sustainable Cities and Communities

• While the focus is on a university, a large institution like Harvard functions like a small community and its waste management practices have a significant local environmental impact. By reducing its food waste, Harvard is contributing to better overall waste management in its community. The article mentions this is part of a “broader institutional commitment to health, cultural, religious, and environmental sensitivities,” which aligns with creating more sustainable and resilient communities.

2. What specific targets under those SDGs can be identified based on the article’s content?

Under SDG 12: Responsible Consumption and Production

• Target 12.3: By 2030, halve per capita global food waste at the retail and consumer levels and reduce food losses along production and supply chains. The article is entirely focused on this target. Harvard’s initiative aims to reduce “pre-consumer waste” by identifying and quantifying “food scraps” to prevent overproduction. The goal is to “minimize waste” directly at the production level in the kitchens.
• Target 12.5: By 2030, substantially reduce waste generation through prevention, reduction, recycling and reuse. The project’s core strategy is waste prevention. The article explains how the data from Winnow allows the culinary team to “adjust production records and minimize waste.” This is a clear example of waste reduction through prevention, as opposed to recycling or disposal.

Under SDG 9: Industry, Innovation, and Infrastructure

• Target 9.4: By 2030, upgrade infrastructure and retrofit industries to make them sustainable… The implementation of the Winnow AI system is a clear example of upgrading infrastructure to achieve sustainability. The article notes that the dining staff previously relied on “handwritten, ‘cumbersome’ production records,” and the new technology makes it “so easy and much more accurate to gather information,” thus retrofitting their operational processes for sustainability.

Under SDG 11: Sustainable Cities and Communities

• Target 11.6: By 2030, reduce the adverse per capita environmental impact of cities, including by paying special attention to air quality and municipal and other waste management. Harvard’s effort to “mature our own systems and practices for effectiveness and efficiency” in managing food waste directly contributes to this target by reducing the amount of waste the institution sends to municipal systems (compost).

3. Are there any indicators mentioned or implied in the article that can be used to measure progress towards the identified targets?

• Quantitative Data on Food Waste: The article explicitly states that the Winnow system uses a “floor scale and a smart camera to weigh food waste in the compost bin” and produces “very good, accurate data.” This weight of food waste is a direct indicator used to measure the volume of waste being generated and, consequently, the reduction achieved.
• Specific Reduction Goal: A clear, measurable indicator is mentioned in the article. Smith Haneef, the HUDS managing director, states, “The hard internal goal that we’ve set is could we achieve 15 percent food waste production [reduction] at the kitchen production level.” This percentage serves as a specific performance indicator for the project’s success.
• Identification of Overproduced Items: The data from the system provides specific insights that act as indicators of inefficiency. The article notes, “The number one [overproduced item] actually is rice… We saw it across the three locations and that was a very easy fix — like immediate.” Tracking the top wasted items is an indicator used to guide corrective actions.
• Menu Cycle Efficiency: The article implies an operational indicator for waste reduction by mentioning the change in menu cycles. The staff “shortened the menu cycles from four weeks to three,” noting that “the three week cycle menu is actually more efficient” and helps reduce waste by limiting the variety of items in inventory.

4. Create a table with three columns titled ‘SDGs, Targets and Indicators” to present the findings from analyzing the article.

Source: thecrimson.com