Can AI Write about Craft Beer?

Artificial intelligence (AI) seems to be all over our news, our social media, even in some of our jobs. Predictions of its effects on our economy and society range from moderated utopian to science-fiction apocalyptic, with the reality likely to be found somewhere in the middle. So let’s ask the question for this space: “Can AI write about craft beer?”

My professional background is in software, so I probably have a better understanding of the technology than the average craft beer consumer; however, I’m not an expert in this area by any means. I have had the opportunity to work with these tools in my real-world, pay-the-bills desk job, so I have more than a casual experience with each. But I will leave the technical details of tokens, transformers, large language models and the ethics of training to other sources, and here deal strictly with the results of an experiment any of us could replicate.

The premise is simple: Feed the same prompt into a few commercial AI packages and judge the results. Each response has been abbreviated in the interest of space and making the point; the full text is largely irrelevant.

“Write an original paragraph about the state of the current craft beer market in North Texas.”

From Gemini (Google‘s AI assistant), we get:

The North Texas craft beer market is navigating a complex period of maturation, shifting away from rapid, footprint-expanding production and moving toward highly localized, taproom-centric sustainability. Following broader national shifts in consumer preferences that demand unique, super-premium profiles alongside lighter, health-conscious alternatives... (122 words)

From ChatGPT by OpenAI, one of the earliest AI products, we get:

The craft beer market in North Texas is in a period of maturation rather than expansion. After years of rapid growth, local breweries are facing tighter competition, rising production costs, and changing consumer preferences that increasingly favor lighter, lower-alcohol beverages, hard seltzers, and nonalcoholic options... (126 words)

From Anthropic‘s Claude, currently one of the leading AI tools:

The North Texas craft beer scene is navigating a period of resilience and recalibration. Overall production in the region dipped about 3% in 2025, a more modest decline than the 6% drop recorded the year prior — and actually outperforming the national craft beer volume decrease of 4% reported by the Brewers Association... (244 words)

Each of these tools did a perfectly adequate job of delivering a cohesive (if rather flat) summary as directed by the prompt. Claude seemed to take the extra step of sprinkling in some measured statistics and mentioning by name a couple of local North Texas breweries, which is appreciated over the somewhat generic text from the other two. Significantly, both Gemini and ChatGPT did list their sources, ranging from Beer Connoisseur articles, content from websites BeerAdvocate and Beer in Big D, and (oddly) a preprint of an unpublished paper from a faculty member of Kharkiv National University in Ukraine.

These results beg the question at the core debate about AI: Did these software tools actually “write” anything original, or are they just summary algorithms for other people’s work accessed online? Certainly, the words strung together are new compositions (in as far as plagiarism is avoided) but they are still based heavily on original content from other authors. One could argue that this is the nature of writing at a fundamental level, and that this practice is no different than what thousands of bloggers do every day. But more on that in a moment.

Let’s step up the game with another, more pointed query:

“Write an original review of Mosaic IPA from Community Beer Company of Dallas, TX.”

Gemini and ChatGPT produced very serviceable content, single paragraphs of largely boilerplate with descriptive buzzwords gleaned from beer review sites and the company website. Overachiever that it is, Claude gave us five full paragraphs with an actual rating score at the end:

In a craft beer landscape crowded with hazy New England-style IPAs and aggressively bitter West Coast offerings, Community Beer Company's Mosaic IPA has always charted its own course—and it continues to hold up as one of the most dependable IPAs in North Texas. Clocking in at a robust 8.6% ABV, this American-style IPA pours a warm light amber... 

...It's a bold, unapologetic IPA that balances ambition with drinkability, and it remains a proud marker of what Dallas craft brewing can produce. 4/5

An impressive review, to say the least.

However—and this is a huge but for the argument at hand—Claude has never tasted Community‘s Mosaic IPA, or any IPA, or any beer at all. It is a disembodied program lacking a body, a mouth, tastebuds or a brain to be fogged with intoxicating beverages. ChatGPT has never visited a single craft brewery in North Texas (or anywhere else), and Gemini has not had conversations with professional brewers about managing supply shortages or navigating pandemic restrictions. None of these products have spent years watching new breweries being built and struggling breweries close their doors. None have memory or experience outside the online session used to pose the question, and will discard the entire exchange once the window is closed.

As good as these AI models can be, and they can be frighteningly good in some respects, they are obligate consumers of other material. Original becomes a blurry concept, as this type of software is nothing but a rules engine for its programmed logic and has no utility until it is “trained” by digesting volumes of published content by human authors, both contemporary and public domain. In biological terms, this software is a commensal organism with regard to human-generated content, like a barnacle on a whale. Its output is (likely) not incorrect and (hopefully) not misleading, but it is at best an abridgement of other works.

Like restaurants, national parks and live performances, craft beer is experiential. Consumers can discuss the objective business aspects and cultural impacts, but they also spend hours sampling, comparing, analyzing and rating their individual interpretations of flavors and aromas and mouthfeel. Consumers interact with other consumers, with brewery employees, with retailers, and form their own opinions from their personal perspectives. People take “beer vacations,” traveling across the country or around the world to visit specific breweries or beer festivals and happily document their stays on social media.

Writing about craft beer is more than reporting objective facts and analyses. It is conveying a personal experience with a beverage, and through that product insight into a subculture that spans the nation if not the globe. And for that, a physical presence with biological senses and awareness is mandatory, and AI has not yet achieved that—if it ever will. PH

* Actual photo of the author working in his office. Totally not AI-generated at all.