Preventing Food Recalls Using Artificial Intelligence

By Lois Harris

Artificial intelligence (AI) can provide food manufacturing companies with a much better chance of avoiding recalls, according to Harjeet Bajaj.

“If you can predict an adverse situation related to food quality, you have a chance of preventing it from actually happening,” the President and CEO of Savormetrics says, adding that preventive quality assurance (QA) and quality control (QC) is much less costly than reacting after the fact, in terms of people’s health, the firm’s bottom line and its reputation.

Bajaj’s Mississauga, Ontario-based company makes ‘FoodSafe Analyzers’, fully integrated tools for analysing food quality that uses AI and machine learning.

Systems that run on AI are much more accurate than humans can ever be partly because they can measure and correlate thousands more features than we can. They’re also much faster. The AI-driven sensors can provide real-time feedback right in the plant, rather than having to wait three to four days for laboratory results.

These tools are extremely flexible because they can be customized to fit the manufacturer’s or retailer’s needs. Regulatory and company QA standards can be loaded into the software, which can even provide direction to employees on the next steps if it detects an issue.

“We call them Zappers – basically the device has sensors that, when passed over the product, can provide you with relevant quality metrics,” he says, noting as examples the shelf life of produce or taste profile of processed food.

He says that employee training for the devices only takes about an hour, and the system setup only requires a minimal amount of samples because AI can extract features and draw correlations that are far beyond what humans can.

“Using AI, you can extract and correlate hundreds of thousands of features instantaneously to attain a highly accurate value – something that is impossible to do with human-generated mathematical models,” he says, explaining the precision of the measurements that can be had with AI.

“This system can even pick up chemicals and predict chemical reactions that may lead to carcinogen formations at the parts per billion level,” he says.

Bajaj says that Savormetrics provides customers with a six-month payback period to get a good return on investment.

About the Author:
Lois Harris is principal writer and editor at Wordswork Communications, and she has been part of GFSR’s editorial community since 2017. She brings a strong background in writing for a number of food manufacturers, non-profit organizations, governments and universities to her work on GFSR’s behalf, and has a keen interest in the food safety issues that affect consumers and industry.