So here is what little I understand. Elon Musk’s X (not his son, but his social media enterprise), which I am never on, has an affiliated AI chatbox called Grok (spelled, perhaps, with an invisible x), which I have never been near.

I think that the difference between a search engine and a chatbox is that a search engine gives you information and references, while a chatbox “chats” with you. Sort of like when you go to a store’s website and see a little box that says: type in your question and a representative will answer you and help you out.

So Grok is a chatbox and I think you can type in a question like “Who is the second best person in the history of the world?” and you get an answer like “That’s an easy one, Arthur. Adolf Hitler”.

Then you might ask why that is and get an answer like: “He knew how to handle the Jewish problem”.

And on and on.

Now, you don’t get an answer because a 300-pound man is sitting in his parents’ basement smoking weed, listening to heavy metal, and responding to random queries. You get answers because a very smart fellow named Al Go-Rithm is looking at all the data he has collected (and purloined) and stored in data centers in, who knows, Manassas, Kazakhstan and Lapland, and coming out with the best answer he can for your precise question.

If you asked this same question through any competing service, it is doubtful that you would get Adolf Hitler as your answer.

In fact, I went to ChatGPT and asked the same question, and was told it depended on my values.  If I was religiously oriented, it could be Buddha (the assumption was I would have put Jesus in furst place). If I were scientifically oriented, it could be Newton or Einstein (the other might be first). Hitler was nowhere to be found.

So what makes Grok different? Ready for this? As I understand it, Al Go-Rithm uses an additional source that other AI platforms properly ignore. Grok adds in all the comments people you and me type into or forward in X. These random entries become part of Grok’s database.

But, as they say, that’s not all. Grok gives, within all the material ut has collected from X users, special prominence to material circulated by Elon the Magnificent himself. Yes, all the garbage transmitted by Musk into X (and we have seen story after story about some of that) gets fed into the Grok data grinder.

And Elon, who gave a Heil Donald (or maybe a Screw Donald, Heil Me) salute on Inauguration Day, has put some positive info about Nazis more than once into his feed.

On X, Elon has over 200,000,000 followers (yes, you read that correctly). What he writes on forwards there is very important and potentially very dangerous. But at least on X, you know that what you are reading is something that Elon has selected, and it represents Elon’s subjective thinking. You can then consider him a god or a fool. That is up to you.

But Grok is a different kind of animal. Grok is not expected to reflect the subjective thinking of Elon Musk. It is expected to be the results of Al Go-Rithm’s data mining. It is expected to be as factual and objective as possible, and to tell you when objectivity is not possible.

So, Grok has apologized and made some changes to its processing. Or so it says.

But, for those of us who worry about AI’s influence on social thinking, maybe we have actually undervalued its dangers. What if (to take antisemitism as an example), the most powerful AI platforms throught the world began spewing out antisemitic tropes as fact. Jews have a secret cabal of men who meet to conspire to take over the world. Jews kill Christian children because their blood is needed to bake matzohs. The Holocaust never ever happened.

There is no control over what goes into these data bases. Each platform sets its own rules on how their data bases are mined and how answers to questions are developed.

We see here today how a large section of the American people can be fooled by a charismatic cliwn to their detriment and that of the country and the world.

The Fascist (IMHO) threat facing us today from Trump and his followers is great. What if AI is coopted by like-minded people, and none of the facts it spits out are factual. Where are we then?