In March, Japanese telecommunications company NEC partnered with the craft brewery Coedo Brewery to analyze beer flavors preferred by different generations using artificial intelligence (AI), developing four craft beers tailored for those in their 20s, 30s, 40s, and 50s. The beers created by AI received mostly positive feedback during tasting events.
AI has now entered the brewery. It learns all the components that affect the taste of alcohol and designs the fermentation process to create new flavors that consumers want. Beyond reflecting generational preferences, it is expected that personalized beers will be produced in the future.
◇AI introduced at a brewery with a 150-year history
In celebration of its 150th anniversary in 2023, Beck's, a beer company from Bremen, Germany, hired a new brewing artisan—an interactive AI called ChatGPT. The company asked ChatGPT to develop new beer recipes using only hops, yeast, water, and malt. As a result, a lager beer with a subtle sweetness, bitterness from the hops, and rich foam was born, named 'Beck's Autonomous.' People reacted positively, saying it was better than existing lagers.
Beck's is not the only company that has entrusted taste creation to AI. In 2023, Atwater Brewery in the U.S. introduced an India Pale Ale (IPA) with a refreshing flavor designed by AI. Prinz Pinakatt, the head of the beer business at Tilray Brands, Atwater's parent company, noted on the 27th in The Economist that "AI technology provides new recipes that were previously unthinkable." St Austell Brewery in the UK also developed a Tropical IPA named 'robot-made craft beer' using AI.
As AI is introduced into beer development, the related industry is also growing. A representative example is brewing enzymes. Global market research firm Research and Markets predicted in its '2025 Brewing Enzymes Market Analysis Report' released on the 28th that the enzyme market, valued at $671.45 million (approximately 9.373 billion won) in 2025, will grow by 7.2% annually, reaching $952.45 million (approximately 13.296 trillion won) by 2030.
Recently, gluten-free beers have emerged for consumers who cannot digest wheat, and low-calorie beers with reduced carbohydrates as well as non-alcoholic beers are also gaining popularity. Enzymes are essential to the development of such beers. For example, amylase breaks down carbohydrates in grains to aid fermentation by yeast.
The report stated, "In particular, the popularity of craft beer is rising in India, Japan, and Australia, with local ingredients such as rice and sorghum being used, making enzymes like amylase, protease, and beta-glucanase essential supplies," and added, "The integration of enzyme biotechnology with AI and digital brewing technologies such as IoT (Internet of Things) supports market growth."
◇Learning about beer flavor components and tasting information
The introduction of AI by beer companies has secured scientific evidence. The Drink Business, a British alcohol trade magazine, reported on the 28th that "last year scientists found a method to train AI to identify the compounds that give beer its flavor," stating that "now it has become possible to predict the beers consumers will like the most and to further enhance the flavors."
Kevin Verstrepen, a professor at the University of Leuven in Belgium, announced in Nature Communications last March that he analyzed the compounds that give beer its flavor using AI and developed methods for improving taste.
Like other foods, beer contains hundreds of aromatic molecules. People detect these molecules through their tongue and nose and perceive them as a single taste in the brain. A panel of 16 experts scored 250 types of beer sold in Belgium over three years based on 50 characteristics such as flavor, sweetness, and acidity. The research team also collected 180,000 consumer reviews online.
The research team identified over 200 compounds involved in creating beer flavor. Esters produced by yeast and terpenoids released from hops are representative examples. These contribute fruity notes to the beer. The flavor of beer is also affected by alcohol content, acidity (pH), and sweetness.
The research team also fed data on people's evaluations of beer flavors and flavor components into the AI. The AI, through machine learning, found the optimal conditions for producing beer flavors independently. As a result, it predicted that adding specific components such as lactic acid or glycerol to commercial beers would yield better tastes.
Indeed, beers made following AI's advice received better evaluations, whether they contained alcohol or were non-alcoholic. The research team recently focused on the rapidly increasing consumption of non-alcoholic beers. Consumers who want to drink beer but are wary of alcohol are on the rise, but many still evaluate the flavors as not matching real beer. Professor Verstrepen stated, "AI has the greatest potential to improve the taste of non-alcoholic beer."
References
Nature Communications (2024), DOI: https://doi.org/10.1038/s41467-024-46346-0