FDA Moves into Second Phase of AI Imported Seafood Pilot Program
FDA starts the second pilot for AI based market surveillance. In a world full of data and limited resources, this is the agency’s attempt to make more targeted efforts to combat food safety violations.
For us it is just another proof point that AI is becoming increasingly a part of food safety.
Why start with seafood?
It is not a coincidence that FDA starts with seafood. With 35,129 records the fishery / seafood category represents the largest food category of import refusals, covering 19% of the total. This is also because according to FDA “the United States imports upwards of 94 percent of its seafood supply.”
Following seafood, we can expect the next category of AI program to be on vegetables / vegetable products, which is the second biggest product category.
FDA Import Refusals 2010 - 2021 (Feb.9th) Source: FDA
How does it work?
Machine learning is a subsection of artificial intelligence. According to FDA, “machine learning is a type of AI that makes it possible to rapidly analyze data, automatically identifying connections and patterns in data that people or even the agency’s current rules-based screening system cannot see.”
This explanation fits also very well with the iComplai machine learning / prediction systems. Here you can see more detailed information.
What is the impact on importers / exporters worldwide?
Machine learning algorithms are able to predict the risky areas. The predictions can include risks in terms of suppliers, country of origin, product and substance.
By understanding the machine learning algorithms, the importers / exporters of not only of seafood but also of other categories can better focus on the risks that are more likely to happen and intensify the efforts in these areas.
Are predictions always the same?
No, predictions are like “living organisms” and they may change over time. As new data is fed into the system, the algorithm will run again, recalculate the patterns and connections and make new predictions based on that.
What is the impact AI on food safety?
With more targeted approach, it would be possible to have more targeted checks and find a higher percentage of violations in the overall shipments within the same effort.
This goes not only for FDA surveillance, but can also be applied to all food manufacturing / processing facilities worldwide.
FDA is one of the first agencies to integrate it in their surveillance activities but for sure they will not be the last.
See the press release of FDA here.
To know more about how you can benefit from AI for food safety, contact us today.