There have been swift advances in the world of artificial intelligence (AI) recently, and the emergence of tools like ChatGPT have made this rapidly-evolving technology more accessible to the public than ever before. As well as helping to answer some of the trickiest questions in fields such as science and medicine, AI is also playing an important role in the food and drink industry, with applications ranging from creating new recipes for dishes, to identifying the origin of wines.
These new ways of using AI are helping those in the industry to respond to culinary trends more quickly than ever before, providing customers with new dishes that excite and satisfy them, both in restaurants and supermarkets. But perhaps the most valuable use of this technology in the industry is the part it looks set to play in combatting the counterfeit wine trade.
A sophisticated palette: how AI can identify the origin of wines
Machine learning and data analytics are now being used to train AI so it can successfully identify the origin of wines, specific to the chateaux and vineyards in which they have been created. By analysing data patterns and chemical compositions, the technology can pinpoint exactly where some of the most valuable wines on the market have originated.
A new study from the University of Geneva has produced impressive results. By using a machine to vaporise 80 red wines from Bordeaux in France, researchers were able to separate each product into a unique set of chemical components. When this data was fed into the AI, its algorithm correctly stated with 100% accuracy where each batch of wine it was tested on was made. The AI could even correctly plot the co-ordinates of each wine’s location on a map, with the data matching perfectly with the exact estates where all the individual wines originated.
At this stage, the AI did struggle to identify what year the product was created, only guessing correctly approximately 50% of the time – a testament to the intricacy of the winemaking process. But it’s likely the accuracy of this will improve as the technology becomes more sophisticated – and the results on location in themselves have big implications for the industry.
Why is this useful to industry professionals?
For the first time, winemakers now have concrete proof that each wine has its own ‘terroir’ – otherwise known as an individual profile taste – giving it a unique flavour, solely based on the vineyard in which it is produced. Factors such as soil, topography, and growth habitat are all part of the terroir, and until now, this unshakable belief held by industry professionals has been the subject of debate. Now, AI is providing scientific backing to support the existence of terroirs, and the unique properties they bestow on each wine.
As satisfying as this AI-enabled “we told you so” will be, these developments in AI technology could also be used for what is perhaps an even more important purpose: combatting the counterfeit wine trade. It’s estimated that Europe’s bootleg and counterfeit alcohol markets cost genuine traders roughly £2.5 billion a year. Using AI technology, wine could soon be routinely tested to prove whether or not the contents match what the label proclaims it to be, enabling traders to know for sure that they are buying or selling the real thing, and stopping the counterfeit wine industry in its tracks.
Alexandre Pouget, a lead researcher on the University of Geneva project, explains: “There’s a lot of wine fraud around, with people making up some crap in their garage, printing off labels, and selling it for thousands of dollars. We show for the first time that we have enough sensitivity with our chemical techniques to tell the difference. Our results show that it is possible to identify the geographical origin of a wine with 100% accuracy, by applying dimensionality reduction techniques to gas chromatograms. We’re at the basic science stage right now. To turn this into a practical application, I’m saying two to four years of work.”
AI’s applications in the beverage industry are not just limited to wine. AI has also been used to detect the production techniques and regional twists of certain spirits, including deciphering the quality of specific whisky products and their backgrounds. In the beer industry, AI can note the hops and malts that create certain beers, and even predict flavour profiles based on factors such as brewing methods and locations. And AI tech can go a step further with cocktails, where algorithms can be used to create new and exciting recipes for mixologists.
Cooking up a storm: AI and the food industry
In the same way that AI has the potential to improve our beverages, the technology is also being used to spice up our culinary lives, suggesting bold combinations to elevate the dishes served at restaurants as well as those sold in supermarkets. AI-inspired cooking is thriving, with McKinsey’s 2022 Global Survey on AI finding that its use in the food industry has doubled since 2017. UK supermarket brand Waitrose provides one example of this – it has recently deployed AI to create new dishes for its shelves, using data from menus, online resources, and social media posts to concoct recipes for its Japanese food range in response to the latest food trends.
Restaurants are also turning to AI to create new culinary delights for their customers, including a US taco chain who used ChatGPT to create its “Chat GPTaco”, which swiftly became its best-selling and best-rated taco of 2023. And AI is also helping with day-to-day restaurant management, where its applications include automating booking processes, stock management, planning and forecasting, and personalising the experience for repeat customers.
The major part now played by AI technology in developments across the food and drink industry is often met with mixed feelings. Food and drink are strongly linked to human emotion, and used as vehicles for connection and bringing people together as social beings. The work of our great chefs is considered a very human and creative craft, and some question whether the use of AI in the creation process has a detrimental effect by diluting this human connection.
However, with AI’s potential to play a part in eliminating counterfeit goods trades in sectors ranging from luxury fashion to wine, it’s clear that striking the balance between maintaining the art and human touch in the products we consume and produce, and ensuring creators do not lose out to fraudsters, will enable AI to play a positive role in the future.