Sweet or sour? AI powered device achieves human-like sense of taste

Graphene-based sensor is a step towards restoring taste for those with neurological conditions

Researchers are one step closer to developing a device that can replicate our sense of taste.Credit: timsa/Getty
A device that combines a graphene-derived material with machine learning to detect salty, bitter, sweet and sour flavours could one day help to restore a lost sense of taste to people with neurological conditions.
The system, described in Proceedings of the National Academy of Sciences on 7 July, builds on previous work using atomically thin layers of carbon to detect flavours. Unlike earlier models, it can operate in wet conditions similar to those of the human mouth.
Although further work is needed before the device can be used in the clinic, “it will be a huge success if we can give an individual back some taste experiences”, says Andrea Lavazza, a neuroethicist at Pegaso University in Naples, Italy, who was not involved in the work.
Partial or complete loss of taste can occur as a result of conditions such as Parkinson’s disease or multiple sclerosis, or during the aftermath of a stroke.
Graphene ‘tongue’
Scientists have long recognized the potential of graphene — which consists of a 2D layer of carbon atoms arranged in a hexagonal lattice — as a chemical sensor, because of its electrical conductivity and sensitivity to different molecules.
In the latest study, the researchers built a taste sensor using layers of graphene oxide — a material derived from graphene, which has a similar structure and properties. Like graphene, graphene oxide has a large surface area and experiences changes in electrical conductivity when it interacts with different chemicals. These variations can be measured, then interpreted using machine learning to distinguish between tastes and build a ‘memory’ of flavours — in the same way our brains process nerve signals generated by taste receptors on our tongue when they are triggered by different chemicals.
Using several sample chemicals (including salty sodium chloride and bitter magnesium sulfate) the researchers ‘trained’ the device to perceive flavours associated with certain chemical structures. It could then recognize these flavours in chemicals it hadn’t encountered, with an accuracy of nearly 90%. It could even be taught to recognize more complex flavours, such as coffee and cola.
Lavazza says the development is “a very, very significant step” towards restoring taste for those with neurological conditions, although he emphasizes that real taste perception is a complex sense that is influenced by smell and texture.
Login or create a free account to read this content
Gain free access to this article, as well as selected content from this journal and more on nature.com
or
Sign in or create an accountdoi: https://doi.org/10.1038/d41586-025-02158-w
This story originally appeared on: Nature - Author:Katie Kavanagh