Can AI beat a Master of Wine in a Wine Tasting? Master vs Machine

Master of wine and the machines are Coming for all of us ever since Chad GPT Was launched by public use people are Running around screaming tearing their Hair out because artificial intelligence Is going to change our lives I myself Wasn't too scared because which computer Would ever be able to do this tough job Sitting around tasting wine all day However I was recently contacted by one Of my subscribers who challenged me to a Tasting battle because he has developed An AI wine taster so well the machine Beat the master or not we find out Artificial intelligence or machine Learning are the buzzwords of the Year Already and people are wondering how This new technology will change the way We work communicate and live our lives I'm one of the over 100 million active Users on chat GPT it appears to be the Leading AI platform at the moment and it Can produce really interesting results Even though it sometimes makes mistakes Constantine biomaster of wine is a wine Educator consultant and master of wine Based in Germany on his YouTube channel He shares his expertise on wine by Producing educational videos I asked Chpt to explain what my channel is about And the results were really interesting There were some mistakes there I don't Really interview wine makers on this Channel but the results were flattering

Enough so that I accept some errors Chatgpt can also write pretty good Tasting notes it obviously doesn't taste The wine instead it searches the net for Information on that wine and then it Builds a note based on patterns that it Has been trained on as an AI language Model I don't have personal taste Preferences and I don't have access to The latest sensory data to evaluate the Quality of wine at least it's more Honest than most tasters about its Tasting ability but as you can see in The example it gets things wrong the 2004 pictures never scored 96 points on Robert and the words It Quotes to describe the wine do not Appear on the website nevertheless this Is interesting technology and when Rudolph zillion contacted me about his AI wine taster I was all ears his system Is supposed to produce objective neutral Unbiased tasting notes so it doesn't Really tell you whether a wine is good Or bad that's really up to you it works Different to chat GPT because the sample Of the wine is also being analyzed and The information on the Alcohol level's Acidity and sugar are being used in Order to improve the results so I tasted Some wines a few weeks ago and then sent The samples to a lab in Sweden and I Just received the results but first of All let Constantine from the past

Explain to you what exactly happened [Music] This is Constantine from the past Talking to you I've set up the blind Testing already I don't really know Whether I should do it blind but anyways I'm just going to do it blind I've Poured all four wines using the kurava Because I want to be able to re-taste The wines when I get the results from The lab A bottle of wine from each of these Bottles will be sent to the lab for Analysis and for for creating the AI Response to the wine and I think now I'm Just going to taste them when one has a Pale yellow color with a slight Touch of Green on the nose it's super aromatic There's quite a lot of Gooseberry some Cassis some Yeah passion fruit flavors coming Through which points me pretty much Directly to sauvignon blanc on the pad It's actually quite rich with good Freshness but not a very pronounced Acidity I think this tastes like German Sauvignon blanc and but I'd be in like Ryan Hassan or the false with this wine It's quite a well-made wine it's not Extraordinary but it's pretty good so Let's check whether I'm right so this is Wine number one It was on the screw cap so that's why There's a corava capsule on top and it

Is a sauvignon blanc from the falz 2021 Vintage 12 of alcohol Pretty good wine let's move on so wine Number two is quite a bit darker in Color it's a little bit more golden Tinge there a few cork pieces floating Around in the glass I don't know how That happened through the korava but Yeah it's a little bit more profound in Color on the nose it's actually far less Aromatic so there's more notes of ripe Apple a little bit of lemon but there's Not this exotic expressive fruit that One one brought to the table on the part It's actually quite rich and Concentrated there's a bit more alcohol There I think so I'd say this is Probably around 13 and a half percent of Alcohol and the acidity is not to Pronounce which kind of points me to a Warmer region not a heart region but a But a warmer region for me the smells of Weissbergunda Pinot Silvana it could Also be Chardonnay for example but I'm More torn between Pinot Blanc and Sylvana but on the palette this kind Concentration and richness is more Consistent with Vice bugunda and I'd say It's probably from a warmer area in a Cool region like Germany I actually Don't think it's as good as the previous One I think it's a little bit well it's A little bit too made there's also some Banana flavor coming through it doesn't

Necessarily appear to me too much but Anyways let's check what what it Actually is damn I forgot about Karina Vetlina this is Karina vetlina which Well I at least sometimes throw in the Same bucket as Silvana and vice paguna It's not super aromatic it can be quite Concentrated and this is exactly exactly That so there's quite a bit of Concentration there it's actually 14.5 Percent of alcohol so it's very rich and Big and well for me that's just a little Bit too much so wine 3 has a dark purple Core gets a little lighter to the rim But it's quite concentrated and rich in Terms of color on the nose it smells of Blackberries blueberries there's also a Little bit of pepperiness coming through Little rusticity so there's also some Herbaceous notes some flavors of black Tea on the palette it's actually quite Intense the the tenons are grippy quite A lot of body there I would say maybe 14 Of alcohol the acidity is rather low and Well all in all it's a pretty good wine I would actually go straight for Surah Here I think this combination of ripe Fruit flavors and the black peppery Notes coming through really remind me of The profile of Surah not necessarily From France from the Rhone where it's Where it's basically at home I would Actually go for South Africa here it Definitely has quite a lot of bite and

Grip and I feel like suras from South Africa tend to be a little bit more Structured Let's check All right It is Surah from South Africa And I got a South African wine right so That needs to go down in the history Books it's from Molina a very famous Producer in swotland in South Africa It's 13.5 percent of alcohol and well It's it's it's a pretty good wine it Actually says five on the back but this Is wide four I still haven't found my Bag number four and this one is actually Pretty light in color I can kind of see My hands through it so it's a yeah it's A light purple color on the nose It smells of ripe cherries a little bit Of strawberries and there's also some Black tea component coming through and Apparent it's actually quite juicy Refreshing fine tenons but they are not Very pronounced so they are more like Low to medium and well the Finish is Fresh and Lively for me this melts and Tastes very much of the old world could Be or maybe Barbara Barbara tends to bring quite a lot of Cherry food flavored or chetu tends to Be a little bit rounder and this has Quite a bit of acidity so I'm going this Is a Barbara a pretty good one but not a Great one for sure it's more of an

Entry-level fresh and fruity style yeah I go here I would say is also around 13 And a half and let's see what it is Damn It's Lange nepiolo so Well I'm not completely wrong but I'm also not really right but now let's See How the machine did [Music] And what the machine actually tells us About these wines these results were not Tampered with as far as I understand so It's basically just data in and then Results out by the AI wine taster so Well let's see what it came up with so I'm starting off with the sauvignon Blanc from the files we locked in all of The data what grape variety it was how It was made all that was entered by Rudolph and then the expected Aromas That came out that the machine told us Were basically Citrus Gooseberry green Apples blackcurrant leaves Nettles and Asparagus pretty obvious results for Sauvignon blonde so it doesn't Necessarily take super intelligent Machine to tell us that it also told us Some things about the taste perception The wine has some Taste of sweetness Fullness is not apparent in this wine Light-buddied food Aromas and tastes are Clearly noticed the wine feels soft but With some hints of bitterness the wine

Will have a long shelf life probably at Least 10 years I think that doesn't really make a lot Of sense I wouldn't say that this Sauvignon law should be aged for 10 Years maybe it could survive this long But this is really a wine made for Immediate consumption the results for The gruna are pretty similar it lists of A whole range of different flavors that You often find in Una Felina and it also Says that the shelf life is at least 10 Years I think this is also not a one That's supposed to be aged this long I Mean maybe a little bit longer than the Sauvignon blonde but not 10 years but Then again shelf life and maturity Window are not the same thing but most Wines don't go bad so there's no point In kind of putting a best before date Onto it but yeah it will survive the Next 10 years it's probably not a wine That you should age for 10 years so the Next one was the Malino Surah and it was A really good Surah I thought Um well flavor wise again you can find All of those hurray flavors listed here On top of that you get some notes on the Aging in Barrel because the wine was Aged in Barrel but the machine knew that Beforehand so it doesn't take a genius To put in Oak and vanilla flavor Um well when it comes to the body and The fruitiness the astringency the

Sweetness all that makes sense the Sweetness is very low at 2 out of a Hundred but the sugar content is also at 0.8 according to the analysis so that Makes a lot of sense the system again Tells us that the shelf life is 10 years And in this case I think it's actually Right but I kind of get the feeling that It always tells us that wine has a shelf Life of 10 years but let's see what it Said about the next one so for the Nebulo it actually lists some flavors That I find a little bit odd I mean There's Cherry raspberry rose hip Rose Violet all those flavors that you often Find in nebiolo white pepper not Necessarily but well anyways the body The fruitiness the astringency Um pretty much spot on and the sweetness At two makes a little sense because it Only has 0.2 grams of sugar per liter And Yeah it only has 20 milligrams of Sulfides per liter which is fairly low For a not natural wine I would say the Description of the taste makes sense Mostly and The Taste will be rather dry For sure fullness is not clearly Noticeable light bodied wine I don't Really understand how it comes up with That conclusion because this is a 14 Acro wine and it actually knows that Because this kind of data was entered Into the system

Um the tenants are clearly detectable on The palette for sure and the shelf life Is again 10 years so am I worried about My job after this session not really I Think when people want to learn about Wine they really want to learn about Wine from another person so I don't Think anyone will ever kind of think Well I want to know what chat GPT thinks About this wine however I think this Could improve the quality of information That is available out there it obviously Has to be checked and edited in order to Make sure that there's not wrong Information in circulation but this Could for example work in a supermarket Where they don't necessarily have a wine Expert on staff that they could gather Quite a lot of information from these Documents in order to train their staff To improve the quality of the Information on the shelves or maybe on Their websites it could also serve as a Basis of information for copywriters who Write about wine but apart from that I Don't think it serves a lot of purposes Yet it's in the name machine learning so As the machine learns more and more it Will improve its results so I think We're still pretty much early days and At some point this could be maybe Something that you could just pick up And put on your website as a description Of a wine but then again

I wouldn't necessarily want that I Wouldn't necessarily like someone who Sells wine to use a machine to create Tasting notes but maybe that's just me I Want to thank Rudolph for helping me Putting together this video I've linked Up his website in the description and I Added some information on him so check That out if you want to learn more so Thank you for watching if you like this Video then please like it down here Subscribe to my channel if you haven't Done so already a question of the day is Is there a future for AI and wine what Do you think let me know Down Below in The comments I hope I see you guys again Soon until then stay thirsty Thank you