New technologies against online fraud

Ai (Artificial Intelligence) to defeat international fraud in the world of wine and beyond.

This is the aim of a joint research project between the Central Inspectorate for the Protection of Quality and Fraud Repression of Agricultural and Food Products (Icqrf) and the Inter-University Centre for Security and Crime Sciences of the Universities of Trento and Verona. Verifood is based on the application of state-of-the-art artificial intelligence techniques to criminological analyses and on the contribution of cognitive scientists who translate the Icqrf inspectors' thinking and methodologies into rules for computer scientists. The aim is to cover an increasing number of protected geographical indications, with the cooperation of the Consorzi di tutela (protection consortia), and also to identify the most effective online frauds, i.e. those most capable of deceiving potential victims, and thus the ads most likely to succeed. After a year of work, the prototype is already at an advanced stage. Over the next two years, the system will be calibrated, continually refined and expanded to cover more and more geographical indications and to improve its ability to recognise misleading ads, including through images.

While waiting for the project to come to fruition, artificial intelligence is already being used to predict plant diseases, optimise grape production and manage agricultural resources more efficiently. Thelearning machine, for example, allows predictive analysis of the behaviour of certain genes in stress situations, and helps predict, for example, which ones are most important for plant survival in the face of the cold weather, which this year in the second half of April, affected the flowering of the vines, as technicians from Veneto Agricoltura found.

Or, through information technology, data are acquired to optimise wine production in terms of quality and quantity, as was the case with 'Mappiamo', an experiment in which Cantina Valpolicella Negrar, in the Veronese region, and the cooperative wineries La Guardiense (Benevento) and Citra (Chieti) participated. Winegrower members were also involved, in order to develop models and forecasting algorithms capable of outlining trends in productivity and grape quality vineyard by vineyard.