“It’s a bit of a know-it-all” says Raoul Masangcay, founder of Elias Wicked Ales & Spirits “but the good thing is it never gets drunk or has hangovers so it’s always available to brainstorm ideas.”
That was his team’s experience of running a collab with the latest internet sensation, ChatGPT, an artificial intelligence (AI) chatbot developed by OpenAI that launched in November 2022
The Philippine brewery which brews out of Quezon City, Metro Manila, decided for their latest seasonal beer release to cook up a collaboration brew with the AI chatbot.
The recipe for their ‘Foggy Daze Pale Ale’ was formulated almost entirely by AI algorithms, using data from thousands of different beers to determine the perfect combination of ingredients and brewing techniques to create a delicious, one-of-a-kind brew.
The collaboration seemed to the team an exciting way to explore new possibilities and push the boundaries of what’s possible. By working together they hoped to leverage the strengths of both human creativity and AI’s ability to analyse vast amounts of data to create some new and interesting flavour profiles.
The AI algorithm behind Foggy Daze (ChatGPT also came up with the name) is based on a machine learning model that, once set the task, would comb through the massive database of online beer recipes, ingredient lists, and flavor profiles and create what it determined to be the best possible concoction given the request parameters.
Raoul had asked ChatGPT, “Can you make a hazy pale ale recipe for 250 litres that uses pale malt, wheat malt and American hops”?
“Sure thing!” the AI replied, and almost instantly set itself to printing out the requested recipe.
ChatGPT’s original Hazy Pale Ale recipe (Image: Elias Wicked Ales & Spirits)
The algorithm was able to identify the key characteristics of a good hazy pale ale, such as the ‘perfect’ balance of hops and malts, the ideal brewing temperature and duration, and the optimal fermentation process.
Once the AI had generated a recipe for Foggy Daze, the brewers used their combined experience to make some slight adjustments and fine-tune it to suit the real-world brewing environment.
The malt volumes, hopping schedule and brewing temperature were slightly adjusted to meet the team’s target SG/IBU’s. As a courtesy to ChatGPT they informed it of the changes to which it warned “using less Cascade hops may result in a milder bitterness and flavor compared to using the original amount”. They thanked it for it’s astute observation and proceeded with their changes.
The Elias Wicked Ales team are delighted with the result of this collaboration between AI and humans. A golden hazy pale ale with a pillowy white head and a delicious balance of citrus, tropical and pine hop characters and slightly sweet and bready malty notes. The addition of the wheat leads to a delightfully creamy mouthfeel that makes it incredibly easy to drink.
The final hop schedule included both early and late additions of cryo hops, resulting in a clean bitterness and intense flavour and aroma that doesn’t overwhelm the malt characters.
The collab result – a smooth, creamy hazy pale ale “Foggy Daze” (Photo: Elias Wicked Ales)
As AI technology continues to advance, Raoul feels that we can definitely expect to see more collaborations like this one in the future, where machines and humans work together to create new products and drinking experiences.
“AI definitely has the potential to be a valuable tool for recipe creation in the future” he observed. “With its ability to analyze vast amounts of data and identify patterns and relationships it can help brewers identify new flavor combinations and optimize recipes based on a variety of factors such as ingredients, brewing process, and customer preferences.”
However he does caution that he doesn’t think AI will completely replace human creativity and expertise in recipe creation.
“Brewing is an art as much as it is a science, and there are certain aspects of recipe development that require intuition, experience, and a deep understanding of the brewing process. AI can help with the technical aspects of recipe creation, but it cannot replicate the unique creativity and personal touch that humans bring to the craft.”
Looking beyond the brewhouse there are many other areas where AI has the potential to help breweries develop their businesses.
AI’s could help predict consumer demand, identifying optimal inventory levels and predict shipping times to help supply chain optimization. They could also be used to help reduce waste, improve efficiency, and lower costs.
And how about in marketing? Consumer insights, trends and preferences can all be analysed at the click of a button to help breweries better target their marketing efforts and create new products that appeal to their target audience.
We will surely see more use of AI in the brewing industry in the coming years, but for now it seems brewers are content to use it as a convenient, if somewhat lazy brewing partner – after all ChatGPT isn’t about to throw on a pair of gum boots and haul grain with you.
The Elias Wicked Ales brew team continue to push flavour boundaries from their brewery in Quezon City, Philippines
