The Food and Drug Administration (FDA) reported an increased detection rate of up to 3.8% using the AI-powered Border Prediction Intelligent System for inspecting imported fruits. This system, operational since 2010, reduces manpower and inspection costs by over US$123,000 annually.
FDA Deputy Director-General Lin Chin-fu explained that the system analyzes cloud-based data from domestic and international sources, including product inspection records and global food safety alerts. It builds predictive models using big data analysis to decide if a batch of agricultural products should undergo inspection.
Imported agricultural products must initially go through the Import Food Information System. The Border Prediction Intelligent System evaluates them to determine if sampling is necessary, testing for pesticide residues, heavy metals, food additives, and microbial content.
Imports into Taiwan have grown from 400,000 batches in 2011 to 760,000 recently. The FDA emphasizes AI's role given limited resources, expanding product variety, and increasing import volume. The agency plans to increase inspectors at the border from 71 to over 100.
Lin highlighted enhanced inspections of Japanese strawberries with a high rate of failed safety checks since June 2023. In late 2022, 317 batches were inspected, with 2.2% failing safety standards. Early 2024 inspections found apples with pesticides and heavy metals above regulations.
In January 2024, banned Sudan red dye was detected in Chinese chili powder in Yunlin. Subsequent detections in February and March resulted in 22 additional batches identified. Importers face fines exceeding US$61,500.
Source: Taiwan News