Is ChatGPT a sommelier in the making?

ChatGPT scored an impressive 92% on the introductory Court of Master Sommelier test. Are human sommeliers concerned?|

Are wine professionals intimidated by ChatGPT, that upstart chatbot that recently passed three Master Sommelier theory exams?

Chatbots are software designed to interact conversationally with humans. Think of the customer service pop-ups you see when you log onto your cell phone account.

Launched in November 2022, ChatGPT is trained to respond to questions with a detailed answer or offer a solution.

“GPT-4 is a large multimodal model (accepting image and text inputs, emitting text outputs) that, while less capable than humans in many real-world scenarios, exhibits human-level performance on various professional and academic benchmarks,” according to OpenAI, which developed the software and launched it in partnership with Microsoft.

The rigorous Master Sommelier exam is split into theory and practical sections. Wine professionals study for months to pass it.

The questions are detailed, covering alternative methods of production for sparkling wines, principal sparkling wines of European countries and grape varietals and types of wines used in sparkling wine production in major wine countries, for example.

ChatGPT showed its chops when it scored 92% on the introductory Court of Master Sommelier test, 86% on the certified Sommelier exam and 77% on the Advanced Sommelier exam, according to drinksbusiness.com.

It’s worth noting that while ChatGPT has had great success with the theory-based portions of the sommelier exams, it cannot fully pass without acing the tasting section.

Alex Sarovich, executive wine director at Healdsburg’s Little Saint, said she’s not losing sleep over the wine prowess of ChatGPT.

“It goes to show the exam isn’t trying to be tricky,” she said. “If you study the information, anyone can pass. So that’s a good thing.”

Sarovich doesn’t think ChatGPT will replace sommeliers or assume some of their duties.

“I don’t think there’s a robot in the world that can replace a meaningful human interaction,” she said. “I try to stay positive. We could go down that robot rabbit hole, but I’d rather not.”

On one hand I agree with Sarovich. To date, no bot, chatbot or robot can replace a meaningful human interaction. Wine professionals, on the contrary, are truly gifted when it comes to meaningful human interactions about wine. Most found their calling in the industry after tasting a compelling wine. When they tasted it, they were amazed at how striking a wine can be and how it can be a springboard to talk about history, culture, art, aesthetics and sensibility.

In short, these professionals want others to find wine just as bewitching as they do.

On the other hand, the people programming chatbots are using unique methods to help them gain sensibilities. For example, a robot, clearly not equipped with eyesight, can find a red truck in a picture. How? Because a robot can interpret pixels.

With this understanding, I’m guessing that robots — who are clearly not equipped with a palate — may be able to evaluate wine someday using analytical tools of some sort.

ChatGPT may be just an extension of a search engine like Google, a wealth of knowledge for us to access. Who would complain about technology that improves our lives with ease and efficiency? Think Uber, iPhone, Apps, Apple Watch.

But then again, what if artificial intelligence continues to program senses? While it may seem a reach that some iteration of ChatGPT could develop a palate, I’m not convinced of anything these days.

So, as we all talk about ChatGPT as though it were a precocious child, I’m always left wondering if programmers will ever teach it to have a great palate.

And if you find all this disturbing, you can always strike up a conversation with ChatGPT about it.

You can reach Wine Writer Peg Melnik at 707-521-5310 or peg.melnik@pressdemocrat.com.

UPDATED: Please read and follow our commenting policy:
  • This is a family newspaper, please use a kind and respectful tone.
  • No profanity, hate speech or personal attacks. No off-topic remarks.
  • No disinformation about current events.
  • We will remove any comments — or commenters — that do not follow this commenting policy.