This electronic tongue is constructed from graphene, a single layer of carbon atoms arranged in a honeycomb structure, which is integrated with an AI neural network.

To replicate human taste perception, the researchers modeled the device after the gustatory cortex of the brain, enabling it to process and interpret complex flavors.

Scientists at Pennsylvania State University have developed an innovative 'electronic tongue' that utilizes artificial intelligence to assess food safety and freshness.

The neural network initially achieved 91% accuracy in determining liquid acidity, improving to over 95% when it was allowed to define its own parameters.

The device employs an ion-sensitive field-effect transistor to detect chemical ions, converting this data into electrical signals for further analysis.

The AI's ability to account for variations enhances the robustness of the sensor, addressing limitations found in traditional ion-sensitive field-effect transistors.

This technology is poised to serve as a cost-effective solution for various chemical sensing applications across the food supply chain and other sectors.

Its unique sensors are capable of detecting food freshness and harmful contaminants, including PFAS, known as 'forever chemicals' that pose health risks.

The sensor functions like a human tongue, with the artificial neural network mimicking the gustatory cortex, which enhances its ability to evaluate food freshness and identify dangerous contaminants.

Tests have shown that the electronic tongue can distinguish between similar beverages, assess milk dilution, identify spoiled juice, and detect harmful PFAS in water.

The researchers report an overall accuracy of 80% for the device's findings, achieving 98% accuracy in identifying types of fruit juice and 99% accuracy in determining the age of fruit.

The research was reported in the journal Nature, showcasing the intersection of AI and food safety, and emphasizing the challenges of monitoring food freshness due to complex chemical compositions.