Recommendations for the Development of Artificial Intelligence Applications for the Retail Level.

Abstract

Some of the early applications of artificial intelligence (AI) for food safety appear to be intended for use at the level of manufacturing and distribution. Artificial intelligence applications to facilitate foodborne illness outbreak investigations, development of HACCP plans, and food safety root cause analyses at the retail level are needed. For example, the interview form in the International Association for Food Protection booklet, Procedures to Investigate Foodborne Illness, could be filled out by humans, but much of the rest of the forms could be completed by artificial intelligence applications. Humans would still have to do the environmental assessments. Most AI applications to date have consisted of pattern identification. Pattern recognition applications may not be capable of assisting in all the proposed retail applications, but it would not be helpful to propose these retail applications without offering a possible path forward. Progress in the proposed directions may require the development of more robust artificial intelligence based on cognitive models. Because this paradigm shift is less familiar to food safety professionals, a comparison between pattern recognition algorithms and cognitive models is offered. An explanation of cognitive models is included to raise awareness of this approach.