Food banking may be fueled by tangible, non-artificial things – like food and relationships. But it turns out that there is also plenty of room for artificial intelligence to play a role in the fundamental act of feeding people.
As with any organization, the functions of a food bank include a variety of administrative, routine and repetitive tasks. While no AI bot is capable of the manual work of physically moving and distributing food, numerous food banking operations could be enhanced by a shift to AI.
Signs that AI is percolating throughout hunger relief are evident across the nation. One specialty of AI is its ability to quickly analyze lots of data, allowing hunger relief entities to better manage their inventory or make food-purchasing decisions, for example. It can synthesize information in seconds, perhaps to match just-rescued food with pantries that can distribute it. Food Bank News has written about pantries using AI to create training videos and perform real-time translations during conversations with clients. See here for more examples of how AI might be used in food banking.
Virtually every department of a food bank can be helped with AI. “Anytime I see repetitive tasks, I think ‘Okay, a system can do this,’” said Ion Nemteanu, Director of Data, Analytics and AI at Jacobs & Cushman San Diego Food Bank (pictured above).
The goal is to free people up to perform more meaningful tasks. “We want to put the brains of these people who are really mission-driven, compassionate people toward more strategic thinking,” he said.
Nemteanu, who has decades of experience as a data scientist and analyst, started at Jacobs & Cushman San Diego Food Bank about six months ago with the goal of deploying data, technology and AI to help the food bank achieve its mission. He is taking a holistic approach to AI, by seeking to have it integrated into the food bank’s overall technology.
Under Nemteanu’s “buy first, build second” approach, the organization is seeking off-the-shelf tools that already have AI components built in. That includes everything from marketing software that generates social media content to procurement tools that can read purchase orders.
Nemteanu sees a host of mission-driven possibilities, but cautioned against getting caught up in AI “hype.” He said, “Make sure you go into it with a very specific organizational goal, not a technical one. Whether it’s driving up productivity, decreasing waste, or optimizing inventories, identify the problem first. That way, if you apply a solution, you can tangibly see a before and after.”
In New York City, The Roundtable, a network of emergency food providers led by the West Side Campaign Against Hunger (WSCAH), has long had a mission to share food-purchasing data to identify better prices. Now it has an AI tool to help it.
In 2018, The Roundtable and food system consultants KK&P began tracking food prices to help inform bulk purchasing decisions and identify pricing discrepancies between its nine member organizations. This allowed Roundtable members to negotiate better prices with vendors, and switch sources if there were better deals for specific items elsewhere.
However, the actual process of manually reviewing invoices and entering the data into a ledger was painstaking. “You’re squinting at pink carbon paper slips that are difficult for a human to read since they’ve been scanned twice and emailed to you,” said Maxwell Bernstein, a senior analyst at KK&P.
After five years of manually tracking prices, the Roundtable and KK&P decided to pursue a more innovative solution. Working with a software developer, the team customized a proprietary piece of AI software that can much more efficiently scrape price data from scanned invoices and input this data into a spreadsheet. Dubbed the General Receipt Analysis and Invoice Ledger, or GRAIL, the system has had an immediate impact.
Since they started tracking prices, Bernstein said the team logged 4,500 individual rows of data. This took five years. With GRAIL, the Roundtable scraped 8,000 new purchases in just 10 months. “We basically quadrupled the actual size of the data set in less than a year,” Bernstein said.
To date, The Roundtable has spent around $100,000 (provided largely through foundation grants) to develop GRAIL, and anticipates spending roughly that amount again this calendar year. The bulk-purchasing strategy that GRAIL supports provides annual savings that adds up quickly. Bernstein said that the nine organizations that make up the collaborative saved about $600,000 last year by purchasing in bulk rather than by case.
While GRAIL is a relatively new addition to The Roundtable’s approach, it’s already paying dividends. “The job of the data is not to tell anybody what to do,” Bernstein said. “It’s to inform their decision-making and to liberate them from having to hunt down the best possible price on every item every time.”
In New Jersey, Christine Yoon, Director of Operations at Meeting Essential Needs with Dignity (MEND*), is starting small with AI, using standard chatbots to help her answer basic questions, like ‘How many cases of bananas should I buy to feed 150 families?’ or ‘What size truck requires a CDL license?’ “At baseline, everyone should be using the chatbot version of AI,” she said.
MEND also uses AI to automate office tasks, like turning historical data or survey results into presentations or reports. “All that tedious data entry and coding work can be done in minutes,” she said. Looking forward, Yoon would like to have an “agentic workforce” that could carry out more complex tasks like recommendations on what to purchase for specific pantries based on current prices, MEND’s budget, and what’s already in inventory.
In Indianapolis, the Indy Hunger Network added AI to a proprietary program to help it rescue more food that would have been destined for landfills. Indy Hunger Network operates a program called Food Drop, which connects truck drivers carrying rejected but perfectly edible food with local food banks and hunger relief organizations. Initially, this process was completely manual, with truck drivers calling the network, which would then spend hours calling local food banks to identify a drop point for the rejected loads.
Local AI researchers and academics at Purdue University knew that there had to be a better way. They worked with the Indy Hunger Network to develop an AI tool that would automate this process, making it both more efficient and less time-consuming, according to Alex Psomas, Assistant Professor of Computer Science at Purdue University and an AI optimization specialist.
Explaining the system, Psomas said that truck drivers now fill out a form that was designed with the help of Indy Hunger Network. Then, an algorithm creates a list of potential food banks to contact, and automatically sends a message to those contacts via text. “This is a full automation of their system,” Psomas said.
He noted numerous benefits of the automation. “Number one, the person who used to do the work doesn’t have to anymore, so we free up their time.” The second advantage is data. “We have a better idea of how things are happening. If you look at a given match, we can tell precisely and algorithmically what the process was that led to a decision of who gets the food delivery.”
In effect, the AI system allows for better “fair resource allocation,” Psomas said, ensuring that food is distributed equitably to hunger relief organizations instead of just the first food bank that happened to answer the phone.
The benefits of improved allocation and time savings come at a negligible cost, Psomas noted. “We’re talking under $10 a month.”
Both Psomas and Nemteanu pointed to a massive untapped resource of researchers, grad students and others in academia who would be “very happy to do these things as projects,” Psomas said. Nemteanu noted that food banks with a bit more means could consider hiring outside AI specialists or in-house technical staff to help increase efficiency and reduce costs in the long term. – Mike Peterson
Mike Peterson is a writer, editor, and media strategist based in San Diego, California.
* Editor Chris Costanzo is a board member of MEND.
PHOTO, TOP: Ion Nemteanu, a data scientist, was recently hired to be the Director of Data, Analytics and AI at Jacobs & Cushman San Diego Food Bank.
Like what you’re reading?
Support Food Bank News
This article was made possible by the readers who support Food Bank News, a national, editorially independent, nonprofit media organization. Food Bank News is not funded by any government agencies, nor is it part of a larger association or corporation. Your support helps ensure our continued solutions-oriented coverage of best practices in hunger relief. Thank you!
