Actually, this has been a good fortnight for the media, with regard to Artificial Intelligence (AI), hasn’t it? Several groups of the originators of this form of computing have come out of the shadows, warning all of us about the potential dangers of letting commercial companies develop this sort of computer activity unsupervised. Two good introductory discussions include these:
Way back in 2018, I published this blog post : Artificial intelligence in the wine industry? Not yet, please! Since then (2020), I have asked: Should we worry about Artificial Intelligence in wine writing? In this current blog post, I will return to the overall topic, by writing about some of the potential general problems of AI, just in case the wine industry tries to get too involved without understanding what they are doing (eg. How AI will be a game-changer for the wine industry).
The whole idea of AI is precisely what its name says — trying to develop a computer version of natural intelligence. The idea is that, if our own brain cells can be intelligent, then so can computer circuits. The basic issue is how to actually go about arranging this.
One common approach these days is called “machine learning”. The basic idea here is that we give the computer massive amounts of what we think is information relevant to some topic, then let the computer troll through it all, looking for common patterns. The computer then uses these patterns to produce apparently “intelligent” responses to future situations, by regurgitating what it deems to be the most likely response. Somehow, this is meant to mimic what human beings learn to do as they grow up.
Unfortunately, this is not quite how we do it, as there is a basic contradiction here. AI simply finds a consensus in a massive block of data, which comes from the past; but this does not necessarily say anything about the future. So, one part of the process is talking about changing things for the future, while the other is about automating access to information from the past. Maybe these are connected, and maybe they are not — this is what we are doing when we say: “I will have to think about it.”
Let’s start by looking at a couple of the absurdities that can be produced by this computerized process. My erstwhile correspondent, Bob Henry, once tried asking Bard about me and my blog. Bard is described by its creator, Google, as a “conversational AI model” (“a machine-learning breakthrough”, no less!). The response Bob got from Bard was this:
David Morrison is an American astronomer, a senior scientist at the Solar System Exploration Research Virtual Institute, at NASA Ames Research Center in Mountain View, California. He is the former director of the Carl Sagan Center for Study of Life in the Universe at the SETI Institute and of the NASA Lunar Science Institute. He is the past Director of Space at NASA Ames. Morrison is credited as a founder of the multi-disciplinary field of astrobiology. Morrison is best known for his work in risk assessment of near Earth objects such as asteroids and comets.
The Wine Gourd wine blog is a website that provides information about wine, including reviews, recipes, and tips. The blog was created by David Morrison, a wine enthusiast and educator. The Wine Gourd blog covers a wide range of topics related to wine, from the basics of winemaking to more advanced topics such as wine tasting and pairing. The blog also features reviews of wines from around the world, as well as recipes and tips for pairing wine with food.
The Wine Gourd blog is a valuable resource for anyone interested in learning more about wine. The blog is well-written and informative, and it provides a wealth of information about wine. The Wine Gourd blog is a great place to start if you are new to wine, or if you are looking to learn more about this fascinating beverage.
The Wine Gourd wine blog is a website that provides information about wine, including reviews, recipes, and tips. The blog was created by David Morrison, a wine enthusiast and educator. The Wine Gourd blog covers a wide range of topics related to wine, from the basics of winemaking to more advanced topics such as wine tasting and pairing. The blog also features reviews of wines from around the world, as well as recipes and tips for pairing wine with food.
The Wine Gourd blog is a valuable resource for anyone interested in learning more about wine. The blog is well-written and informative, and it provides a wealth of information about wine. The Wine Gourd blog is a great place to start if you are new to wine, or if you are looking to learn more about this fascinating beverage.
Clearly, the first paragraph is about someone other than me; and the second paragraph is a load of nonsense. I quite like the third paragraph, though! Still, I would not rely too much on Bard, just at the moment, if I was you.
Another absurd example comes from here in Sweden. A national supermarket chain asked ChatGPT (another AI-powered language model, this time developed by OpenAI) to produce a recipe for an Easter cake. All that ChatGPT did was throw together a lot of the foods associated with Easter, as shown in the picture below. This haphazard cake did not go down well with the customers — it was disparaged as “Egg in green grass”.
Actually, it is often ChatGPT that generates the main discussion of problems, or else other chatbots. There is little point in me going into details in this blog post, so here is a list of reading for you:
- The false promise of ChatGPT
- OpenAI’s new ChatGPT bot: 10 dangerous things it’s capable of
- Why a conversation with Bing’s chatbot left me deeply unsettled
- ChatGPT proves that AI still has a racism problem
- OpenAI’s chatbot spits out biased musings, despite guardrails
- AI can be racist, sexist and creepy; what should we do about it?
- What ChatGPT gets right — and wrong — about Washington wine
In addition to all of this stuff about chatbots, there are many other concerns, as well, such as:
I conclude that the answer to my title question does appear to be: Not just at the moment, thankyou.
Now, maybe I am just a Grumpy Old Man, bemoaning that the world is changing around me, at a time when I would like to slide into a quiet old age. However, I have lived most of my adult life throughout the Computer Revolution, and so I’ve seen too much unjustified hype before, and I am therefore not so easily fooled any more.
A suitable analogy here is guns and gunpowder (the latter was introduced to Europe from China). These were used to feed us at first; but so-called “bad actors” quickly learned to use them for attacks on humans, instead. This started off small, but gradually increased, with hordes of horseman sometimes streaming in on placid settlers. Then the guns became cannons; and then the powder became part of bombs. Eventually, we developed nuclear weapons — this is a long way from a few single-shot rifles. So, at that stage we backed off pretty quickly, because not every product of the human mind is necessarily in our own best interest, globally. As a young man I was very glad about that, because the alternative future looked pretty bleak to me, back in the 1970s. Furthermore, we can easily detect when someone detonates an atomic bomb, but we currently have no idea what AI is doing behind the scenes.
Anyway, all that Machine Learning does at the moment is take huge amounts of data, and then searches for patterns in it, which it regurgitates upon request (Can today’s AI truly learn on its own? Not likely). There is more to intelligence than this. To me, the success of chatbots simply shows us how many lonely people there are in the world. Chat with your friends, instead; and if you don't have any, then use your own intelligence to learn how to make some, rather than relying on the artificial kind. However, all of this does not mean that AI is not useful, for many tasks.
Fortunately, at least some politicians are apparently listening (White House says it will look at AI regulations, legislation): their basic principles are said to include the need to “evaluate, verify, and validate the safety, security, and efficacy of AI systems.” We shall see!
- Amazon CEO Andy Jassy says ‘really good’ A.I. models take ‘billions of dollars’ to train
- 14 million jobs worldwide will vanish in the next 5 years, new economic report finds
- Google to launch AI-driven ad campaigns
- What is your winery's origin story? How context can elevate ChatGPT!
- VinterActive introduces AI-assisted text messaging for DTC wine marketing
- Gloria Ferrer vineyards a Sonoma County pioneer in using AI to monitor crops
- The driver-optional EV tractor at center of agtech boom right next to Silicon Valley
- ChatGPT makes wine in south of France
- Democratizing wine investment using AI and blockchain for authentication and valuation
- Here are 34 jobs AI could never take — and how much they pay
I conclude that the answer to my title question does appear to be: Not just at the moment, thankyou.
Now, maybe I am just a Grumpy Old Man, bemoaning that the world is changing around me, at a time when I would like to slide into a quiet old age. However, I have lived most of my adult life throughout the Computer Revolution, and so I’ve seen too much unjustified hype before, and I am therefore not so easily fooled any more.
A suitable analogy here is guns and gunpowder (the latter was introduced to Europe from China). These were used to feed us at first; but so-called “bad actors” quickly learned to use them for attacks on humans, instead. This started off small, but gradually increased, with hordes of horseman sometimes streaming in on placid settlers. Then the guns became cannons; and then the powder became part of bombs. Eventually, we developed nuclear weapons — this is a long way from a few single-shot rifles. So, at that stage we backed off pretty quickly, because not every product of the human mind is necessarily in our own best interest, globally. As a young man I was very glad about that, because the alternative future looked pretty bleak to me, back in the 1970s. Furthermore, we can easily detect when someone detonates an atomic bomb, but we currently have no idea what AI is doing behind the scenes.
Anyway, all that Machine Learning does at the moment is take huge amounts of data, and then searches for patterns in it, which it regurgitates upon request (Can today’s AI truly learn on its own? Not likely). There is more to intelligence than this. To me, the success of chatbots simply shows us how many lonely people there are in the world. Chat with your friends, instead; and if you don't have any, then use your own intelligence to learn how to make some, rather than relying on the artificial kind. However, all of this does not mean that AI is not useful, for many tasks.
Fortunately, at least some politicians are apparently listening (White House says it will look at AI regulations, legislation): their basic principles are said to include the need to “evaluate, verify, and validate the safety, security, and efficacy of AI systems.” We shall see!
No comments:
Post a Comment