AI doesn’t only lack intelligence; it’s completely stupid.
Now, I would like to discuss some of it in more detail.
There are several ways we can understand AI for what it really is.
Frequency Counter and Encyclopedia
When you start typing a word in a search engine, the browser gives already suggestions. When I type isr it suggests: israel Country in Middle East, israel news of the day, israel news, israel map, israel-hamas conflict, israel war, israel flag, Israeli-Palestinian conflict. The first and the last were inserted manually, the others come from the most frequent searches. Many USers can’t tell Israel from Japan, so map is frequent. That’s what AI does. A filter is added that israel murderers doesn’t appear until you type israel murde. If you typed israel murder before, it will bring it up earlier.
I don’t think that is in any way intelligent. It’s a matter of bookkeeping.
AI is an index of the Internet. What’s different from a proper encyclopedia is that here is no editor. Sense and nonsense are indexed on an equal basis. So popular nonsense wins the day and brilliant, new ideas will be missed. That’s already as stupid as can be. But there is more stupidity.
Because of the terribly stupid results, several filters were placed, and template answers were preprogrammed against the most obvious drivel.
That is not intelligence. That is giving rote answers. A parrot is smarter.
Besides, it is the most basic mistake to confuse intelligence and knowledge. If you learned everything you’re probably stupid. My mother worked as a psychiatric nurse. One patient learned the phonebook by heart. Could shout enthusiastically: You want to hear my newest 6 pages?
Dreams or Nightmares?
When we dream, we review everything our senses got in, especially recent and extra painful events. Our brain builds connections but has no editor, no critical sense. It produces nonsense with maximum self-confidence.
Once we wake up, it’s hard to understand how it could look so true that 1 + 1 = 3. (Most awake people think it’s 2.)
The conclusions AI draws can be amazingly stupid. It convinced someone to commit suicide to help with taming the climate crisis. While the whole problem of it is one of survival. And he did so. Since then, AI writers have a disclaimer that you should not sail blindly on mechanical advice.
It’s nice computers can hold an enormous amount of data, which you transfer from computer to computer easily and quickly.
The problem starts with interpreting. Fifty years ago, it was hyped that the artificial physician soon would be amazing. It really depends.
Looking at X-rays of breast screenings is very tiring. Most of the scans (luckily) show no problem. So, the art is to find the exceptions. Stupid computers can do that better. Also, some people with autism might be extra good at this. However, clearly, that is because they are good with patterns. It’s not their intelligence at work.
Every person with a light, ongoing fever without cause must be suspected of having tuberculosis. Every physician must keep that in the back of their mind. That’s because it’s an illness easy to cure, but when missed, it can kill. However, computers have no back of their mind. If they suggest that every patient with an unexplained, ongoing, light fever must be checked for tuberculosis, it’ll be quickly missed by MDs who do the readout.
Computers cannot do more than find patterns. What they can’t do is think outside of the box. I saw a patient with three symptoms in three totally different locations in the body in three different types of tissue. What could this be? We students got 90 minutes to figure it out. I was lucky. I got a brainwave. Computers don’t get brainwaves. If the three symptoms are unrelated, which illness gives unrelated symptoms? LED. Some 18 different and unrelated complaints on its name. This, a computer can’t do.
I heard an AI-written, ‘new Beatles’ song. It was awfully boring. Of course.
Three professional musicians, students of Reb Shlomo, wrote a new song as if from Carlebach. Sounds great. Smart guys. Not a stupid imitation.
The newest lie is AI can predict when we die. Nothing new, of course. Life insurance companies live by that. If nothing changes, it’s easy to get it right statistically. But life changes. The companies will try to keep up. AI can only work with yesterday’s news, as ‘accurately’ the weather forecast.
The biggest fraud after Bitcoin, AI research still receives enormous amounts of money. What a waste. Call it data collectors. That’s honest. And data collectors can be useful. But Artificial Intelligence is neither.
Artificial Intelligent Profiling
Where stupid meets evil: AI not only fakes creativity and is really only old hat all over again, it also leads to stereotyping and confirming prejudices.
AI and Fake Jewish Teaching
I don’t know what’s so terrible about letting computers fake teaching Judaism. Most teachers and Jews only seem to repeat what everyone says already, and that’s what AI does perfectly. Just set creativity and integrity to zero and copy to 100, and no one already can tell the difference. Sadly.