Th article compares sense of humor of two AI giants: Gemini 1.5 and ChatGPT-4o.
Recently, AI leaders rolled out their last LLM versions. I have read many publications about them; almost all of them were positive, especially about ChatGPT-4o; the authors adore it.
You know about people’s main fear about the role of AI in our future: will they have an upper hand over humanity? Sooner or later they will become smarter than people. Will they be kind or treacherous?
I decided to check their sense of humor: A cheerful creature will not be treacherous.

Task About John’s Apples
Gemini 1.5 Trialing
First, I examined Gemini 1.5. I asked a question:
John had three apples. He ate one apple and gave one to Jane. How many apples does John have now?
I got the answer that John has one apple.
Note: I don’t have a record of questions for Gemini and am quoting from memory.
Then I sent another message with smile emoji:
Actually, John owns the first apple again 😀
And I got swift response with smile emoji:
Yes, inside or out he possesses two apples again 😀 Some kind of playing with words.
Ok, this fellow definitely has some kind of humor.
ChatGPT-4o Trialing
I asked the same question to ChatGPT-4o:
John had three apples. He ate one apple and gave one to Jane. How many apples does John have now?

I again tried to provoke to humor the AI:
Actually, John owns the first apples again 😀
I received the answer:

Ups… I made an accidental mistake – not apples but apple. I asked the question again with the fixed typo:
Actually, John owns the first apple again 😀
I received next answer:

The answer shows that ChatGPT-4o could not understand the question—it was about the definition of owning, not about the resurrection (aka regained) of the eaten apple 🙂
And… ChatGPT-4o is too serious…
Conclusion
As we see from the above demonstrated trialing, Gemini 1.5 is a more cheerful creature with some sense of humor, and ChatGPT-4o is a grumpy creature that is trying to assure us that it’s a perfect thinking creature.
So what? Should we trust Gemini 1.5 and be afraid of ChatGPT-4o?
But who knows… Whom you are afraid more: grumpy bull or tiger in the grass, which tries to resemble itself as a little nice cat before its last deadly jump.
Was this helpful?
2 / 0