Food waste in America is a big problem with an estimated 30 - 40 percent of our groceries getting tossed out, costing retailers $57 billion each year. To address this issue, X launched a food waste moonshot in search of a smarter food distribution system. Two and a half years of research and prototyping later, the project is moving up to Google proper for further development at scale.
One of the issues that the X team identified was data siloing within the industry. “Just as food sits in silos across the country, information about food also sits in ‘silos’ in organizations,” Project Delta’s Emily Ma wrote in Medium on Tuesday. “There’s no easy way for food suppliers to share information about their available food, or for food banks to register their needs.”
The X team hopes to recreate the efforts of Kroger and the Southwest Produce Cooperative, which coordinate to route grocery overstock from the supermarkets to food banks in the Tucson, Arizona area, using software. The team has already produced a prototype food distribution system running on Google Cloud, dubbed “Dana-bot” (after Dana Yost, the COO of the Community Food Bank of Southern Arizona). This system, Ma explains, “automatically uploaded information from the Southwest Produce Cooperative’s donated food dataset, categorized and standardized each entry, then matched the food with food banks and food pantries based on real-time needs in the Feeding America network.”
The team has also produced a prototype computer vision system that can help cut down on food waste in commercial kitchens. It essentially watches the trash bin, identifies what is being discarded and tracks how much is getting thrown away in a given time frame. The team set up a 6-month (pre-pandemic) pilot program in 20 Alphabet cafeterias where the cameras collected twice as much data as conventional hand-sorting methods. The team hopes to continue its food identification work once pandemic relents and workers return to Google campuses.