Heard This? Scientists Have Deployed AI To Detect Fraudulent Wines
There is nothing worse than procuring a poor quality wine passed off as premium for a bonafide sommelier. A true one looks for vintages and assesses their authenticity on the basis of the bouquet, texture and flavour. While a stunning Bordeaux can easily be distinguished from a Sauterne by the expert, what if technology can further aid the oenophile to prevent wine frauds which are actually a very lucrative, but of course, illegal trade?
According to a report in The Guardian, scientists have worked with Artificial Intelligence (AI) to create an algorithm that detects fraudulent wines by tracing a drink’s origin with chemical analysis. Wine frauds, as is evident, means a beverage produced in a low-key garage or winery and passed off as premium with use of fancy labels.
The same report alleges that the illicit trade costs European sales nearly $3.2 billion in revenue. Machine learning has been working wonders in several sectors to detect frauds and prevent crimes, and now, winemaking will also be under the AI radar.
How Does The Algorithm Work?
The scientists working on this project use the technique of gas chromatography to analyse 80 wines harvested over 12 years from seven estates in Bordeaux, a prominent wine producing region in France. The Guardian report added that more than distinguishing one individual compound from another, this algorithm is, in fact, drawing on all chemicals detected in the drink to arrive at a reliable signature for each premium liquor. With this, the program creates a grid in which wines with similar signatures are grouped together.
It is possible for researchers to differentiate wines by narrowing down on the subtlest of differences through an analysis of the tiniest chemical compounds using this algorithm. AI will be able to facilitate an analysis of the region in which the wine originated along with the estates on which a particular vino was produced.
What Did The Algorithm Reveal
When this algorithm was actually applied to wines, it showed that there are chemical clusters which correspond to a particular chateaux which are distinct for every different space. This implies AI can distinguish the many tiny molecules in the wine responsible for creating the particular chemical signature of a chateaux.
Moreover, the algorithm showcased how clusters from every different region also revealed the location or the estate from where they emerged, effectively creating a ‘map of Bordeaux’ as the report suggests, based on the chromatography findings.
If that’s not enough, the program is able to scrutinise grapes, soil and microclimate discernible in the chemical compounds of the wine to determine which chateau it originated from. The AI algorithm was able to deliver a near accurate result while tracing the origins of the wine, with just over a 50% success in determining whether it is a vintage.
How Can AI Prevent Wine Frauds
By tracing the origins of the wine, the algorithm can be employed to confirm whether a wine matches the label, making it a very crucial step in Europe where there is a sizable number of fraudulent networks dealing in fake wines.
A scientist working on this project has suggested that along with determining fraudulent wines, the machine learning programme could be used to monitor winemaking quality and technique to ensure that it is well-blended. Since a good Bordeaux is all about right blending which requires quality craftsmanship, AI can actually cut down on costs making it easier and lighter on the pocket to make quality blends.
