An artificial intelligence-based approach to deal with argumentation applied to food quality in a public health policy

Argumentation is a relatively new research area in Artificial Intelligence. Since the early 1980s, its use has been investigated in various frameworks. We propose a general model for recommendation-based argumentation by extending Dung’s seminal argumentation system. This approach is applied to analyse argumentation on food quality in a public health policy. Cereal products, and more specifically bread, are used by decision makers as a healthy lever to fight against diseases such as obesity or diabetes. Our model outputs new recommendations based on stakeholder’s argumentation by targeting some specific audiences.

► We analyze a case study on risk/benefit evaluation in the wheat to bread chain.
► We propose a formal argumentation model for its representation and evaluation.
► Specific recommended actions are computed depending on the considered concerns.
► More complete bread is recommended in the “cardiovascular prevention” scenario.
► Organic, debranned or white flour are recommended for sanitary purposes.