Engineers at the University of Glasgow have recently developed an artificial tongue capable of differentiating between tastes. The tongue technology is based upon a glass wafer featuring three separate processing centres. Each of these is composed of 2m miniature artificial taste buds, just 100 nanometres long – 1,000 times smaller than the width of human hair.
“Essentially, when you shine any given coloured light on the taste buds, they shine another colour back at you,” Dr Alasdair Clark, the lead author of the research from the University of Glasgow, tells Food Spark.
The colour shift, Dr Clark emphasises, is tiny – so tiny that it’s not necessarily something we can see with the human eye. In spite of this, each liquid effectively leaves its own fingerprint – meaning once the AI has sampled multiple types, it can be used to differentiate between tastes.
“This works in a very similar way to the human tongue in that it knows the difference between, say, apple juice and coffee, but it doesn’t know the chemicals that make each substance up,” reveals Dr Clark. “That’s why we’re referring to it as a tongue rather than a traditional sensor.”
According to the engineers, the tongue is capable of recognising differences with greater than 99% accuracy. The technology can easily be incorporated into a small, portable device, according to the researchers, and its usage could range from the environmental monitoring of rivers to the identification or poisons – but what about testing food?
“I’m sure someone cleverer than me could come along and play with it to do food too,” Dr Clark laughs. “The general concept would be the same. It picks up specific chemicals and recognises the difference between item A and B. So there’s the potential for it within food, definitely.”
The IBM approach
The University is far from the first to create an artificial tongue, and another recent project, carried out by IBM Research, also draws inspiration from the way in which humans’ taste.
IBM’s technology uses electrochemical sensors comprised of pairs of electrodes, designed to capture a range of chemical information. The sensors each respond to the presence of a combination of molecules by means of a voltage signal.
Once this ‘fingerprint’ has been established, it is sent to a mobile device, such as a smartphone, where it is transferred to a cloud server. From there, a trained machine learning algorithm compares the digital fingerprint to a database of known liquids. The algorithm identifies which liquids in the database are most chemically similar to the liquid under investigation, and reports back to the mobile app.
IBM’s proof-of-concept suggests the entire process takes less than a minute, and Patrick Ruch, lead investigator of Hypertaste at IBM Research, believes the technology could unlock the door to more cost-effective and immediate alternatives to lab testing.
“In the supply train, at present, once food and drinks are packaged, there is little ability to verify that the package actually contains what is on the label, apart from sending the product to the lab for testing,” Ruch tells Food Spark’s sister title Food Navigator.
“Suppliers acting in bad faith may insert lower-quality products into the supply chain with little risk of getting caught, or counterfeiters may even fake a real product by adding the few analytes which are most likely to be tested for in a lab. This is a real problem with wine, mineral water and even olive oil.”
Looking further afield
The experimentation doesn’t stop there. Washington State University’s very own e-tongue is reportedly a more reliable taster of spicy food than a human.
According to the university’s report, human taste buds become desensitised to spices after only a few samples. Instead, an e-tongue could accurately measure the spiciness of multiple foods for hours at a time.
Earlier this year, the e-tongue was tested on samples of paneer cheese containing different levels of capsaicin, the active compound of chilli peppers. The results were compared to volunteers tasting the same samples, and were proven to more reliably detect capsaicin levels long after the point at which human taste buds “gave out”.
The university believes its e-tongue is also more accomplished at detecting subtle differences in spice. However, it does not believe it would replace testers – only streamline the process.
"[It] would allow testers to narrow a selection down to two or three samples for a human tasting panel if they start from 20 different formulations," researcher Courtney Schlossareck stated in the report. "That would take days to do with people tasting them."
Food Spark has previously reported on tech like the FlavorWiki that could potentially takeover traditional tasting panels.