Key points are not available for this paper at this time.
This study assesses the effectiveness of an intervention employing an AI-based, fully automatic waste-tracking system for food waste reduction in HORECA establishments. Waste-tracking devices were installed in a restaurant within a holiday resort and a business caterer in Germany, a hotel in Switzerland, and two hotels in Greece. The devices utilize computer vision and advanced deep learning algorithms to automatically weigh and optically segregate food waste in real time. At baseline, total food waste was 76.2-121.0 g/meal for the hotels, 99.4 g/meal for the business caterer, and 151.9 g/meal for the restaurant. Avoidable food waste constituted 45 % to73% of the total, attributable to overproduction (20-92 %) and consumers' leftovers (8-80 %). The remaining waste was unavoidable, stemming from preparation procedures (47-99 %) and consumers' leftovers (1-53 %). Vegetables and prepared foods contributed the most to total amounts. This data-driven intervention raised staff awareness towards food waste, facilitating the implementation of corrective actions. Therefore, except for the Swiss hotel that exhibited an increase of 13 %, the intervention was effective in achieving a 23-51 % reduction in food waste, especially in food preparation and overproduction, demonstrating the intervention's transferability across different settings. Additional evidence supported its long-term sustainability. The cost of wasted food per meal was reduced by up to 39 % compared to the baseline. Future studies should explore combining waste-tracking devices with consumer-level interventions to enhance food waste reduction.
Sigala et al. (Sat,) studied this question.