Swiss Researchers Teach AI to Detect Counterfeit Wine

Swiss scientists have developed an artificial intelligence tool designed to identify counterfeit wine, as reported by The Guardian. The AI utilizes gas chromatography, a specialized physicochemical method for separating and identifying compounds within mixtures, to analyze the chemical composition of wine and trace its origin.

To train the AI program, researchers employed gas chromatography, a method commonly used in laboratories to separate and identify components of mixtures. The scientists tested this method on the analysis of 80 wines collected over 12 years from seven different estates in the Bordeaux region of France.

"The first thing we noticed is clusters that correspond to specific chateaus. This immediately told us that there is a chemical fingerprint characteristic of each chateau, regardless of the vintage. It's a general picture of the concentration of many molecules that distinguishes each chateau," explained Professor Alexandre Pujol from the University of Geneva in Switzerland.

Several factors, including grape variety, soil type, microclimate, and the winemaking process, influence the concentration of compounds in the wines of each chateau.

"The power of machine learning for this type of research becomes increasingly apparent with each new application in the food industry and agriculture," said David Jeffery, a wine science lecturer at the University of Adelaide and co-author of the book "Understanding the Chemistry of Wine."

The developed program, worked on by scientists, proved capable of accurately tracing the origin of wine with a precision of up to 99%. However, its ability to differentiate between wine harvests has an accuracy rate of only 50%.

As a reminder, the International Organization of Vine and Wine (OIV) has published its assessment of global wine production in 2023. It is expected that the production volume will be the lowest in the last 60 years, with France leading the industry, according to the OIV announcement.

Based on data from 29 countries that produced 94% of the world's wine in 2022, global wine production (excluding juices and must) in 2023 is estimated to range from 241.7 million hectoliters (hl) to 246.6 million hl, with an average estimate of 244.1 million hl. This represents a 7% decrease compared to the average wine production of the previous year. The projected production this year will be the lowest since 1961 when only 214 million hl were produced, and even lower than the historically low figure in 2017 - 248 million hl.