All in Good Taste

Scientists are attempting to lick food contamination with cutting-edge technology. Say hello to the e-tongue.

By Anne-Frances Hutchinson

Contaminated food poisons 600 million people every year.

The symptoms – which, unfortunately, many of us are much too familiar with – range from mild discomfort to crippling pain. Asking, “Does this taste funny to you?” doesn’t just dare the respondent to play palace taster; foodborne illness is deadly, killing about 420,000 of the world’s people annually.

One in six Americans get sick from foodborne illness each year, resulting in about 128,000 hospitalizations and 3,000 deaths.

Tainted food takes a vicious bite out of corporations and brands, too. Depending on severity, a single outbreak can cost companies anywhere from a few thousand dollars to several millions in lost revenue, lawsuits, fines, higher insurance premiums, and reputational repair, just for starters.

And that doesn’t take food fraud into consideration. Food Safety Net Services defines that “as the act of purposely altering, misrepresenting, mislabeling, substituting, or tampering with any food product at any point along the farm–to–table food supply chain. Fraud can occur in the raw material, in an ingredient, in the final product, or in the food’s packaging.”

Fraudsters mess with food to make money, either by boosting its apparent value or cutting production costs. It’s nothing new, reaching back centuries, but it’s currently on the upswing. Bloomberg estimated that from the first half of 2019 to the same period in 2020, cases rose by 37%.

While the number of U.S. food recalls in 2024 dropped 5% from 2023, serious illnesses and deaths from contaminated food actually rose by 25% during the same period. PIRG’s Food for Thought 2025 report noted that hospitalizations and deaths more than doubled. “Nearly 1,400 people became ill from food they ate in 2024 that was later recalled – 98% of them from just 13 outbreaks,” the authors wrote. “All but one of the 13 outbreaks involved Listeria, Salmonella, or E. coli.”

The Food Safety Modernization Act of 2011 marked a tectonic shift in the way the American public health system approached contamination, shifting from just counting incidents to funding research on new technologies to prevent them. Scientists had been developing food safety technology for many years before that legislation was passed. Finding ways to electronically “taste” food resulted in the creation of an artificial tongue.

Developed in 1995, e-tongue technology delivers precise responses, accurately identifying microbial pathogens, chemical residues, allergens, heavy metals, and other contaminants in liquid-based substances. Sensitive and specific, it is used to authenticate beverages, monitor fermentation processes, and monitor quality control.

However, conventional lab-based testing and manual inspections are slow, labor-intensive, limited in what they can do, and can destroy the samples being analyzed.

Now, researchers at Penn State have given the e-tongue an impressive, futuristic upgrade. By integrating spectral analysis, machine vision, and blockchain into intelligent, integrated platforms, their e-tongue delivers real-time results, continuous monitoring, predictive analytics, and automated decision-making.

Using advanced sensors, AI, machine learning, nanotechnology, and data processing, they created an e-tongue inspired by our gustatory system. The research team included members from Penn State’s Department of Engineering Science and Mechanics and NASA Goddard Space Flight Center and was supported by a grant from NASA.

“We’re trying to make an artificial tongue, but the process of how we experience different foods involves more than just the tongue,” said the study’s corresponding author, Saptarshi Das, the Ackley Professor of Engineering and professor of engineering science and mechanics at Penn State. “We have the tongue itself, consisting of taste receptors that interact with food species and send their information to the gustatory cortex — a biological neural network.”

Our taste buds recognize sweet, sour, bitter, salty, and savory, and over time, our brains train themselves to pick up the subtleties in different flavors. The Penn State team employed deep learning technology to create an artificial neural network that acts like the human brain and adapts to new tastes and sensations.

Study co-author Harikrishnan Ravichandran explained their methodology. “Previously, we investigated how the brain reacts to different tastes and mimicked this process by integrating different 2D materials to develop a kind of blueprint as to how AI can process information more like a human being. Now, in this work, we’re considering several chemicals to see if the sensors can accurately detect them, and furthermore, whether they can detect minute differences between similar foods and discern instances of food safety concerns.”

The science behind the story

The team linked the artificial neural network to a graphene-based conductor that can detect chemical ions. Instead of assigning specific sensors to each different chemical, non-functionalized sensors that detect an entire range of chemicals were used. (The sensors don’t need to be identical to produce results, which may help manufacturers keep costs down.)

They gave the neural network 20 parameters to assess how liquids interacted with the sensor’s electrical properties. Based on these very specific parameters, the e-tongue could accurately detect complex samples, “including watered-down milks, different types of sodas, blends of coffee, and multiple fruit juices at several levels of freshness — and report on their content with greater than 80% accuracy in about a minute,” according to the Graphene Council.

“After achieving a reasonable accuracy with human-selected parameters, we decided to let the neural network define its own figures of merit by providing it with the raw sensor data. We found that the neural network reached a near ideal inference accuracy of more than 95% when utilizing the machine-derived figures of merit rather than the ones provided by humans,” said study co-author Andrew Pannone. Using game theory — a method called Shapley additive explanations — researchers could ask the neural network what it was thinking after it made a decision.

Despite the talents of the e-tongue, tough barriers to adopting AI for food safety abound. Garbage in, garbage out is still the ultimate truth, and AI is only as good as the data it is fed. It is expensive to implement, the algorithms need to be more transparent, and integrations across platforms are problematic. Governance rules must be created and used consistently, and global food safety standards must be established.

Those hurdles don’t bother Das and the team. “We figured out that we can live with imperfection, and that’s what nature is — it’s full of imperfections, but it can still make robust decisions, just like our electronic tongue,” he said.

With its ability to detect subtle changes in freshness, authenticity, and contamination in near-real time, the upgraded e-tongue could become a frontline tool in preventing outbreaks, curbing food fraud, and building greater trust in the global food supply chain. If the technology can be scaled affordably and integrated into existing safety protocols, the next question might not be, “Does this taste funny to you?” but, “Has the e-tongue cleared it?”

—

Food Fighters: AI and the battle to curb food contamination and waste

The e-tongue detects dissolved and liquid substances that human taste buds recognize: salts, sugars, acids, and bitter.  When trained with quality data they can learn to find subtleties and new tastes, just like our own tongues. Unlike human taste receptors, artificial ones are consistent over time.

Traditional food quality assessment methods rely on:

• Human sensory panels
• Chromatography
• Spectrophotometry
• Microbiological tests

Earlier detection systems lacked real-time oversight, took hours or days to get results, and weren’t built for testing outside the lab. AI revolutionizes testing, bringing continuous monitoring, predictive analytics, and automated decision-making throughout the entire food chain.

AI delivers unprecedented results:

96% accuracy finding walnut impurities

96.2% accuracy for meat freshness

93% accuracy in classifying plant diseases

98.4% detection rate of pesticide residues

95.71% identification rates for adulterants

94% capture rate for pathogens within 10 hours¹

96.6% determining spoilage²

15-20% reduction in grain spoilage prediction errors

84% accuracy in linking foods to disease pathways

In food testing, the e-tongue has been used to assess:

• The authenticity of extra virgin olive oil
• Synthetic additives in milk
• Microbial spoilage in meat, poultry, and dairy
• Yogurt and cheese quality
• Real-time fermentation monitoring
• Tea blends and properties
• Fruit juices
• Coffee production methods
• Lead, cadmium, mercury, and other heavy metals

… and too much else to fit into this review!

AI, nanotechnology, and advanced data analysis tools help deliver the goods⁴ with:

• Immediate or near instant substance identification
• Faster sampling speed
• More precise traceability
• Real-time data processing
• Less destructive sampling
• No need to use individual sensors for every sample
• Reduces complexity

Beyond food safety: what else can AI contribute to?

• Personalized nutrition by predicting how diets affect individual health
• Reducing food waste by 14–52% through real-time demand forecasting and inventory management
• Reducing bitterness in pharmaceutical formulations
• Better waste water contamination detection

And now, the easiest pop quiz you’ll take today:

Q: In a blind taste test, what technology can accurately tell the tiny differences between Coke, Pepsi, Diet Coke, and Coke Zero?³

A: Oh, please. After all this? Yes, it’s the e-tongue.

Sources:

1. https://www.azorobotics.com/News.aspx?newsID=16022
2. https://www.psu.edu/news/research/story/matter-taste-electronic-tongue-reveals-ai-inner-thoughts
3. https://nypost.com/2024/10/21/science/ai-tongue-can-taste-the-difference-between-coke-and-pepsi-penn-state-researchers-reveal/
4. https://docs.google.com/document/d/1x_u2pqzP21AdWX_bKXQP6xt0CzJKxknA1WS74cYkGQo/edit?tab=t.0