If self-driving cars can identify objects like other cars or traffic lights, why not use artificial intelligence (AI) to identify harmful bacteria in food.

Turns out that this is not such a far-fetched idea.

“People are using AI to change people lives,” said Luyao Ma, a researcher with the Department of Food Science and Technology at the University of California, Davis.   “As food scientists, we can bring this knowledge into the agricultural food system.”

This optimistic look into the future of food safety is highlighted in a recent report (https://journals.asm.org/doi/10.1128/aem.01828-22) based on research out of the university that explores how artificial intelligence may improve food safety. The lead author of the study, Ma, and her fellow researchers, found that a technique using artificial intelligence and optical imaging can quickly and accurately identify bacteria in food.

Nitin Nitin, professor of food Science and Technology at the university and one of the authors of the report, shares Ma’s optimism about this, saying that with this sort of advancement, “we can deliver an even higher level of food safety.”

“Food safety is an essential part of the food business, so if we can strengthen that area by providing a cost-effective, rapid, highly sensitive and specific approach I think the method, overall, should give consumers more confidence in our food systems as we move forward,” said Nitin.

about AI
Even with all of the buzz about artificial intelligence, most people aren’t sure what it actually is, especially since it’s a rapidly evolving technology. Simply put, it refers to data-laden computer systems that can perform jobs that until just recently only a person could do. This would include reasoning, making decisions, solving problems, recognizing speech, identifying patterns, composing written content, steering a car and analyzing data. In other words, it can use data-laden computer systems to achieve or exceed human intelligence. “Mechanical Intelligence” is one way it has been described.

Some see it as a way to lower operational costs based on what they believe is AI’s potential for greater accuracy in doing repeatable tasks, among them assembling vehicles. Others say it can quickly come up with new content, such as articles, reports, movie scripts, legal briefs, or even images and hit tunes.

But some people fear it will lead to job losses due to increased automation. They also fear that it has the potential to create misinformation and generate biases.

And people ask: Is this the beginning of an AI revolution?

AI and food safety
With advances in artificial intelligence (AI) technologies, the development and implementation of digital food systems are becoming increasingly possible, according to to the University of California Davis report.

The university research involved taking digital images of romaine lettuce, since it has been frequently reported as the vehicle of E. coli outbreaks, with a conventional microscope equipped with white light. From there AI-enabled software was used to detect and identify bacterial micro colonies. Using this method to detect bacteria was found to be a simple, cost-effective and rapid approach compared to the usual ways pathogens are detected.

The results were good. By using this method, referred to as YOLO, they accurately identified 11 out of 12 lettuce samples that were contaminated with E. coli.  

Another definite plus is how quickly the bacteria can be detected and identified. An analysis can be finished within 3 hours.

“This is very rapid screening,” said Ma. “once the food is processed, it can be tested and, on the same day, you can make sure there are no pathogens and distribute it.”

In contrast, standard culture-based detection methods can take 5 to 7 days, by which time the foods likely have been distributed through the supply chain and consumed.

According to the report, early detection of microbial contamination in food is critically important for consumer safety and outbreak prevention. The consumption of contaminated foods was estimated to cause 550 million illnesses (almost 1 in 10 people) and 230,000 deaths worldwide every year.

Ma’s fellow researcher Nintin said that by using YOLO, E. coli could be differentiated from 7 other foodborne bacterial species, including Salmonella, with an average precision of 94 percent.

“There’s the potential to create automated food safety inspections using this AI method,” he said.

Then, too, the methods currently used for most pathogen testing require sophisticated equipment and specialized personnel, which the study says are less accessible to food industries.

Besides health concerns, the early detection of pathogens also significantly reduces the business costs associated with food recalls and liability

Enter the food retailers
The retail food industry employs more than 6 million people and flushes $800 billion into the U.S. economy.

From the producer to the retailer, a strong food safety culture is critical. For the retailer that’s especially true since since retailers are responsible for the food safety of all products sold in their stores.

Bottomline, the retail food industry should prepare for a highly automated future, say food-safety officials.

Yet only 31 percent percent of food retailers use AI, according to a blog put out by FMI, the Food Industry Association. And of those, their main AI applications are leveraging customer data and making decisions on pricing and promotion.

Not a slam dunk; still a way to go
This is definitely an area of interest for the food industry with opportunities in food safety” said Heather Garlich, spokesperson for FMI.

Greg Komar, technology director of the California Leafy Greens Marketing Agreement, shares FMI’s interest in this.

“Our organization is dedicated to improving food safety,” he said.  “Although we’re not using AI now, we are currently investigating it as a tool. By and large we’re looking at any tool that will help us improve food safety in leafy greens.”

Donald Schaffner, scientific editor for the Institute of Food Technologists, said the study on AI and food safety clearly demonstrates the promise of this technology.

“But it is important,” he said, “that we temper the excitement with an awareness of just how far we still have left to go to fully implement this idea.”

And while some of the data contained in the manuscript was generated from micro-colonies on petri plates, which make an ideal model system, Schaffner pointed out that foods are much more complicated.

“ It’s also important to note that this technology is limited to bacteria that are found on surfaces,” he said. “Successfully imaging a micro-colony embedded in a food like hamburger would be much more difficult.”

And while the study does investigate the application of the technology in samplings of Romaine lettuce, and the technology is able to detect relatively low levels of bacteria (e.g., 100 bacteria or 10,000 bacteria), Schaffner said it should be noted that often when pathogenic bacteria are present in foods, they are present at lower levels than this, so important challenges still remain.

Brian Hitchcock, chief science and technology officer at the Institute of Food Technologists, was quick to say that continuous improvement and innovation are critical for food safety in a rapidly evolving food system, which is why “it’s exciting to see innovation at the crossroads of artificial intelligence and microbiology.”

But he also said that at this early stage, it’s difficult to definitively gauge the technology’s potential in the industry. But the prospect of quick and accurate food safety data through this novel approach is undeniably intriguing.

Even so, he pointed out that it’s important to note that more research needs to be conducted to better understand the full capabilities and limitations of AI-based microbiological detection methods in real-world food safety scenarios.

And while AI may potentially aid in preventing outbreaks and illnesses, Hitchcock said it is still important to practice essential food safety practices such as Good Manufacturing Practices, proper cleaning and sanitation, thorough cooking, and regular testing to minimize the risk of harmful bacteria like E. coli in food.”