Sometime in the next couple of years Robert Joseph will buy me a bottle of wine. I believe the erstwhile wine correspondent of the Daily Telegraph to be a man of his word, so I expect he will honour the exact terms of the deal, enter a bricks-and-mortar indy and pay for the bottle with cash. The reason for this purchase will be Robert’s techno-optimism. A few years back, we made a public bet on Twitter on the future of driverless cars. There was one of the flare-ups of techno-babble on the subject, likely caused by the yearly Elon Musk pronouncement that Tesla is less than a year away from autonomous driving. Defeat for me, on the other hand, would involve buying a bottle with cryptocurrency. Amazingly, this was a time when crypto was not considered the prelude to “and then he went to jail”, but a revolution waiting to happen. I am unsure now what it was meant to revolutionise exactly. Maybe the shampoo industry by uploading every flake of dandruff to the blockchain? Something along those lines. The details are hazy.

You probably see where this is going, as we have been living through another techno-babble flare up. The release of ChatGPT has meant that all your favourite epidemiology experts (as seen in the PandemicTM) , have retrained in Artificial Intelligence and can confidently pronounce on how Large Language Models will change the world. The general vibe seems to be that the arrival of chatbots means that writing is passé, school is obsolete, and we will all lose our jobs. It says something about our political predicament that while a hundred years ago people thought technological advances meant less work, today we think they mean fewer jobs.

Now, before you bookmark this to throw it to my face in a decade, my guess is that AI and Natural Language Processing will indeed lead to noticeable changes, mostly for the best. Some jobs will indeed be replaced, and many of them will be in IT. (Murphy’s First Law of Informatics has it that computers exist to help you do things that if computers didn’t exist you wouldn’t have to do in the first place.) It is not a tragedy if a few people with “PHP Ninja” in their LinkedIn profile stop having access to six figure salaries. School will indeed change, as it has been doing for the past two thousand years. It’s just that progress is a linear, not Boolean, function, and things largely progress when and where journalists aren’t looking. Right now, an awful lot of middle-aged people seem to be overexcited by a computer game, playing Twenty Questions with Microsoft’s Clippy.

I spent too much time on futurology, and we are here to talk about wine. As the juice itself seems to be safe, the excitement has been about what these technologies will do to wine communication.  Writing marketing copy is in the firing line, but is that such a bad thing? Churning out the same type of press release and faux tasting notes sounds like a dreary existence. Getting those to be written by a bot sounds ideal; the only thing missing is inventing a bot to read them. Repeat ad infinitum, and we reach marketing telos.

Which brings us to wine critics, the privateers to marketing’s Royal Navy. Could a chatbot replace those? Jay Rayner spent a good 2,000 words pondering this for his restaurant reviews on The Observer, and the closest he came to an insight was that a bot “can’t taste the food”. But restaurant reviewing is multi-faceted. A chemical analysis is not. Is an advanced chatbot armed with a very detailed technical sheet wine criticism’s T-1000?

Anyone that is tempted to answer yes, fundamentally misunderstands what wine criticism is about. People that follow wine critics (and, for the avoidance of doubt, I unequivocally speak here as a follower, not a critic) don’t do so to find out what is “the best” in some objective, mathematic fashion. Granted, part of the attraction is information, which can be automated without too much trouble. But, primarily, we do it to receive aesthetic leadership. We might not even fully realise this is the case. Initially, when we are novices, we largely want to like the things they like. If later we get to know more (and, nota bene, many people that follow critics never fulfil that condition) we are mostly there for the conversation. If we agree or not is almost beside the point. We care about the critic’s opinion the way we care about the opinion of an interesting friend. (The trade can stop shouting; as professionals, they are competitors to critics, not the audience.)

This is the reason why all successful critics have a persona that is larger-than-life. The style can differ: it can be Robert Parker’s aplomb, or Jancis Robinson’s scholarliness, or Tim Atkin’s thoroughness. But there has to be something that is particular and attracts a specific audience. This is why different critics work for different crowds – and also why people with all the qualifications, knowledge, and judging skill in the world fail miserably if their critic persona is colourless. Who wants to consider the opinions of a boring stranger? There is social media for that.

So, I am confident that wine critics will remain immune to chatbots, tastebots, and other technological misfortunes. A chatbot is an excellent text predictor, but hardly a concept inventor. The resulting text predictions might seem miraculous now, but something tells me most kids below 15 will soon develop a knack for telling if something is fake (the Turing test forgot to account for age). A successful wine critic persona just has something very human about it, something that perhaps might be feasible to duplicate – but not invent.

All in all, there might be only one constituency vulnerable to the charms of AI, one group that would gladly let Clippy dictate their drinking. Currently, they pay $80 for some unremarkable Californian Pinot Noir from an estate their favourite tech bro bought six years ago. When a future robot taster uses its superhuman analytical powers to identify Yellowtail Merlot as the optimal wine for them, it will be a win of sorts. See, maybe in my way I am also a techno-optimist.

Photo by Aideal Hwa on Unsplash