Both data collection and analysis are integral to reducing food waste: staff must find out what types of foods are being wasted and in what quantities to then be able to strategize on how to prevent or reduce it. Artificial intelligence (AI) supports those stages of food waste reduction, beginning with the identification of different waste types using machine vision, which becomes increasingly faster and more accurate through machine learning.
“If you literally feed it with thousands of pictures every day (e.g., a thousand pictures of a cucumber), it gets smarter over time,” said Jeroen Feron, growth marketer, Orbisk, which took the COVID-19 downturn as an opportunity to enhance its monitoring technology. “On an international scale, you see a lot of different dishes. For example, in India, we see a lot of different curries, so we have to continuously improve our product [to recognize that type of waste],” he explained.
At the waste analytics stage, AI can make recommendations on processes to reduce food waste based on the data collected in the first phase. “We have large language models that are built into our AI to give quick, actionable recommendations and tips to address waste based on historical trends,” said Jon Garrett, vice president of business development, Leanpath, which uses a cloud-based analytics platform.
Feron believes that food waste management AI will eventually connect to other systems and information sources (e.g., the property management system, purchasing system, weather sites) that will support prediction modeling. For example, the AI might predict that due to the weather, 50 percent fewer guests should be expected at a catered event and recommend that food purchases should be reduced by some degree. “I think there is a lot of ground to win for us in this area,” he said. GS
