Devil’s Advocate: What We Could All Learn From ChatGPT

Robert Joseph thinks about thinking.

Reading time: 3m 20s

Robert Joseph - the Devil's Advocate
Robert Joseph - the Devil's Advocate

Well, what do you know?

Nobody and nothing – not even ChatGPT – knows everything. Like animals, human beings develop awareness and understanding of the stuff that is most relevant to them. This model works very well until they are confronted with circumstances with which they are unfamiliar – when, in the case of animals, a changing climate forces them to look elsewhere for food or brings new predators. Or, for human business professionals, when they encounter regulatory or cultural differences in export markets.

On my travels, I am constantly struck by the chasms in knowledge that persist in the wine industry, despite the instant online availability of encyclopedic levels of information. European understanding of the US market is often surprisingly thin. When I talk about the success of red blends or bourbon-barrel wines or hard seltzers or wine brands changing hands for hundreds of millions of dollars, my words are often met with looks of blank incomprehension. But the same happens in the US when discussions turn to European matters like PIWIs and mandatory nutritional labelling.

We all tend to be most receptive to the kind of information that suits our worldview, irrespective of its source.

Top US wine writers tend to equate ‘big wine’ to American giants like Constellation and Gallo; they rarely include businesses like Grands Chais de France or Felix Solis in Spain with revenues of hundreds of millions of dollars and wine ranges that stretch across their respective countries. Some innocently seem to fondly imagine that most European wine is still largely produced and sold by small family-owned estates – rather than big businesses and the cooperatives that actually process around half of Europe’s annual harvest.

Of course, one cannot separate what any of us think we know from confirmation bias: we all tend to be most receptive to the kind of information that suits our worldview, irrespective of its source. Large numbers of wine professionals seem to ‘know’ that modern consumers increasingly “care about how food and drink are made” and “where they come from”. Or that those same consumers are “increasingly interested in wine” and “looking for autochthonous varieties” and “minerality”.
 

Requests for evidence?

Requests for evidence to support these kinds of statement are rarely fruitful. They’re assertions people who would like to be true believe to be so.

All of which brings me to thinking. Or, more precisely, not thinking about pieces of knowledge that seem to be mutually incompatible. How, for example, does one square the ‘fact’ that consumers are looking for lower-intervention wines with the contradictory ‘fact’ that they are also looking for zero-alcohol wine. What about all the interventionist spinning cones and reverse osmosis machines used to remove the alcohol?

If people really care about where wines are from, why are multi-country brands like I-Heart so successful? And why does Spain sell so much bulk wine to France?

The capacity of thoughtful analysis is still one of the advantages we have over robots and computers, but advantages are of little value if they are not exploited.

My role in writing these columns, is sometimes to give readers new information about stuff they might not previously have thought of interest, and always to try to encourage them to think about what they know, even if that process leads them to disagree with me with even greater certainty.

The capacity of thoughtful analysis is still one of the advantages we have over robots and computers, but advantages are of little value if they are not exploited.

Wine

Traceability is a term that is increasingly used, along with the assertion that "consumers want to know where stuff comes from." A New Zealand company is using technology familiar from TV police forensic programmes to help a super premium wine estate to give assurance about the provenance of its wines.

Reading time: 3m 30s

A conversation with ChatGPT

For the purposes of this column, I asked ChatGPT two questions:

First: “Do you know everything?”

To which it replied:

“I have been trained on a vast amount of data and have access to a wide range of information. However, I don't know everything. There are still many things that I haven't encountered or learned about, and there may be gaps or errors in the data that I have been trained on. Additionally, my responses are generated based on patterns and associations in the data, so they may not always be entirely accurate or complete… In such cases, I rely on feedback from users to improve my performance and minimize any potential biases… I am designed to continuously learn and improve my performance over time.”

And then:

“Do you think that ChatGPT 'thinks'?”

“As an AI language model, I don't ‘think’ in the way that humans do. I don't have consciousness, subjective experiences, or emotions… While my responses may sometimes appear to be similar to human thought processes, they are generated through the use of complex mathematical models and algorithms that identify patterns and associations in the data. I don't have the ability to make decisions or judgments based on personal beliefs, emotions, or preferences…”
 

Lessons to learn?

Whatever you may think of artificial intelligence, perhaps this piece of electronic gadgetry has a few lessons for us all. It acknowledges the gaps in its knowledge, the possibility of being wrong, and a readiness to learn and be corrected. How many of us can honestly say that about ourselves?

Removing personal beliefs, emotions and preferences from the decision-making process is more questionable and could reasonably be thought to be at least slightly psychopathic. But, would checking one’s opinions for confirmation bias while collecting data to support a business decision one wants to make, really be such a bad thing?

Over 2,500 years ago, Confucius said “To know what you know and what you do not know, that is true knowledge.”

So what does that say about ChatGPT?

News

The Italian Wine Crypto Bank (IWCB) has developed a digital sommelier. The system - dubbed ‘Personal GPT Sommelier’ - will be used in the ‘hybrid wine bar’ BG3.0 Winebank in Bergamo. 

Reading time: 1m 30s

 

 

Latest Articles