The human nose and its ability to smell is amazing.

Each nose has approximately 400 odor receptors and is said to be able to detect approximately 1 trillion different odors.

Replicating this level of sensory expertise with scientific instruments is a daunting challenge.

But thanks to recent advances in artificial intelligence (AI), the speed and accuracy levels of modern electronic noses (high-tech sensors that can detect and report specific odors) are rapidly increasing.

Their supporters say they can change food safety.

Common potentially fatal food poisoning bacteria are salmonella and E. coli. Professor Raz Jelinek, co-inventor of the electronic nose called Sensifi and professor of chemistry at Ben-Gurion University in the Negev, Israel, says both of these have their own "electronic personalities." "They have their own electrical signals."

Manufactured by an Israeli company of the same name, the electronic nose contains electrodes coated with carbon nanoparticles. Detects odors and volatile organic compounds (VOCs) emitted by bacteria.

Romaine Lettuce
Image Source, Getty Images
Image Caption,
Contaminated romaine lettuce is said to be one of the biggest causes of food poisoning Different
types of bacteria produce different VOC fingerprints, which vary within the Sensifi machine Generate electrical signals. This is recorded by an AI software system that matches it against a growing database and notifies the user.

Sensifi, launched earlier this year, hopes to change the fight against infectious diseases in the food industry. Currently, most food producers must send samples to laboratories for testing and wait days for the results to come back, said Modi Peled, the company's CEO.

In contrast, Sensifi's e-noses can be used on-site by food companies themselves and are said to provide results within an hour. The company has not disclosed the price of the machine, but says it will be "low cost." The company will instead make most of its revenue from subscription fees.

"Testing methods in the food industry haven't changed for 40 to 50 years," Peled said. “Until now, AI has not really entered this testing area of ​​the market.”

Banners surrounding links to articles about AI
Read additional articles about artificial intelligence

Banner surrounding link to article about AI
Food poisoning remains a serious problem around the world. In the United States, 48 ​​million people, or 1 in 6 people, become ill from food poisoning each year. Of these, 128,000 people have been hospitalized and 3,000 have died.

It has been calculated that 2.4 million cases of food poisoning occur in the UK each year, resulting in an estimated 180 deaths.

"People will say meat, poultry and fish are the main culprits," Peled said. "But the biggest killer of the U.S. food industry over the past five to 10 years is romaine lettuce.

“And the more industrialized the food market is, the more susceptible it is to pathogens.”

German company NTT Data Business Solutions has developed a new way to help train the AI ​​that powers the e-nose it's developing: coffee.

In one test, engineers spent three days placing instant coffee powder next to the AI's sensor. The AI ​​then had to identify one of three options: good coffee, bad coffee (coffee with vinegar), and no coffee at all.

"Odor is not just a gas, it's a unique combination of gases," says Adrian Kostorz, the company's innovation manager. "And there are often variations and very small differences in how people perceive smells."

NTT's sensor is attached to a 3D-printed plastic model of a human nose. The company uses coffee and other foods to train its AI to recognize what it smells like when it's fresh and in good condition. According to the company, this is the "standard odor value."

NTT's sensor is embedded inside a 3D print of a human nose
The idea is to use NTT's e-nose not only to sniff out infectious diseases, but also to monitor the freshness (or not) of food. ) can also be understood. This will allow supermarkets and cafes to know which items to sell without expiration dates first.

"Knowing the odor threshold will help the food industry adapt production, storage, harvesting and processes accordingly," Kostorz added.

But some AI experts say that while the latest e-noses work well, food companies don't expect significant demand due to cost.

“If you're talking about deploying a global network of small detectors from picking to storage to delivery, you need to consider how that will impact your business model,” says the U.S. said Vincent Peters, founder and chief designer of the based company. AI research company Inheritance AI.

"Will the business still exist if we have to implement and maintain that technology? Is supply chain management possible? Will we really implement it? Will it be profitable?
"

Meanwhile, fellow AI expert Kel Carlson of San Francisco-based Domino Data Labs said a motorized nose would require complex fine-tuning for each facility you're working with. “This is a very difficult job in an industry that is not known for adopting new technology,” he says.

But such skepticism doesn't discourage some entrepreneurs.

In New Zealand, a company called Centian Bio announced that it had developed a "biosensor" by copying an insect's antennae. This showed that they were able to clone insect proteins and incorporate them into odor sensors.

As a result of this biotechnology, the company's sensors are "thousands of times more sensitive than a dog's nose," said Andrew Kraliczek, the company's founder and chief technology officer.

He added, "This biosensor-based technology can be used in virtually any application, including food and flavor quality control, food pathogen detection, non-invasive rapid disease diagnosis, sustainable agriculture, and environmental and health monitoring. It can be used anywhere,” he added.