AI Identifies Whiskey Aromas and Origins with High Accuracy

Researchers demonstrate machine learning algorithms that predict whiskey flavor notes and country of origin more consistently than human experts.

A study used two machine learning algorithms to analyze the molecular composition of 16 whiskies, distinguishing between American and Scotch varieties with over 90% accuracy.
The algorithms identified key flavor notes in each whiskey more consistently than a panel of 11 human experts, matching aggregate results from the panel's assessments.
American whiskies were associated with flavors like caramel and compounds such as menthol, while Scotch whiskies were linked to smoky or medicinal notes and compounds like methyl decanoate.
The AI approach could enhance efficiency and consistency in whiskey quality control, ensuring flavor stability across batches and blends.
Researchers suggest potential applications beyond whiskey, including detecting counterfeit products or improving odor management in recycled materials.