Do we really need AI? (9/18/2026)

If you have been following the news, you have probably noticed how prominently artificial intelligence (AI) appears in the headlines. Much of the debate focuses on the data centers AI requires and the communities resisting their expansion. These weeks, some AI experts reportedly warned that AI technology could threaten humanity within the next 10 to 30 years. I wonder if they actually meant human civilization instead humanity in its entirety. But even if isolated indigenous communities in Africa, South America, and Asia survived such a catastrophe, the warning raises a more immediate question: What is AI for, and why do we believe we need it?

Copilot, Gemini, Claude, ChatGPT, and Groc are some of the names we hear about. I have played with all of them in a very limited fashion except for Groc; I refuse to touch things that Musk touches. The AI systems are fun and interesting, I already see it’s evolution over one year. It was a year ago that I asked a few of the AI programs to write my eulogy. I presented the AI eulogies in my men’s group; but did not feel like publishing it here.

To repeat my question: Why do we actually believe we need AI? Yes, at times makes computations easier, it even assists me with complicated questions such as: should I pay off the mortgage on my home, or what is the best route between Yorktown, Virginia and Tyler, Texas avoiding highways as much as possible and driving no more that 200 miles per day? I had an answer to both questions within minutes. I could have figured out the answer to both questions on my own; however, in both cases Claude did it much faster than I could have done it.

So why do we need it? I am afraid that society is getting dumber. The other day I gave the cashier a ten-dollar bill for something that cost $9.85 and he seriously had to examine the screen to see how much cash he needed to return. OK, we do not need AI for that, a simple calculator could figure that one, although he should have paid attention in math classes in school.

Reportedly, AI has now solved one of the remaining algebra problems that challenged researchers. Something we really do not need. The New York times reported that OpenAI disclosed 6 new incidents of “concerning” AI behavior.

How could AI wipe out humanity? All the sudden all airplanes fall out of the sky; boats drift aimlessly on the oceans; all traffic lights stop; electricity disappears; water stops running; hospitals stop functioning; and ever the computers in vehicles stop working. This may not be a quick end; it may take a week or two. Who needs a nuclear Armageddon? Survivalists may hang on a little longer, but even their battery banks have computer chips in them. In an interview Peter Thiel would neither confirm nor deny that the super-rich are expecting or even planning for this (the end of humanity) to happen. Is this why Musk is trying to father a super race? But can you imagine the inbreeding among these kids in future generations? Boy, I need to write a science fiction novel about this scenario.

I wonder if we need a kill switch, in addition to slowing things down? But the AI is spread over multiple computers so killing one may have no effect. Moreover, it would probably be able to replicate itself. Quite an interesting conundrum. But will it really happen or is that disgruntled AI employee just out for publicity, fame and money for his appearance in front of the media or maybe even a future book. Time will tell. All that I know is that we probably can use some checks and balances on AI.

I asked ChatGPT to turn me into a Roman emperor using a regular photograph of my wife and I. Very convincing isn't it?