{"id":9005,"date":"2023-06-12T13:20:40","date_gmt":"2023-06-12T17:20:40","guid":{"rendered":"https:\/\/www.cartoonstock.com\/blog\/?p=9005"},"modified":"2025-12-06T10:15:39","modified_gmt":"2025-12-06T15:15:39","slug":"whats-so-funny-and-why-ai","status":"publish","type":"post","link":"https:\/\/www.cartoonstock.com\/blog\/whats-so-funny-and-why-ai\/","title":{"rendered":"What’s so Funny and Why AI?"},"content":{"rendered":"

First off, I wish everyone would calm down about AI. It’s not going to destroy humanity\u2013 that’s our job. Or, to paraphrase FDR, “Humanity has nothing to fear but humanity itself.”<\/span><\/p>\n

There are real dangers, of course. First and foremost, it could replace cartoonists. But if AI has any real sense, it wouldn’t. Too much work for too little pay. But since it still lacks common sense, it might.<\/span><\/p>\n

If so, I could bear some of the blame. I created The New Yorker caption contest, and the millions of entries to the contest have been used to train it to understand cartoon humor.<\/span><\/p>\n

That said, let’s see how ChatGPT explains the humor of this cartoon of mine which appeared in The New Yorker in 1997 after IBM’s Deep Blue had defeated world chess champion Gary Kasparov, perhaps marking the beginning of the road to our computer overlords lording it over us.<\/span><\/p>\n

\"\"<\/a><\/p>\n

BUY THIS CARTOON<\/a><\/p>\n

ChatGPT Explanation:<\/span><\/p>\n

The humor comes from the absurdity of this situation and the man’s irritation. He doesn’t want his microwave to do anything complex or sophisticated, like playing chess; he just wants it to perform its basic function: reheating his lasagna. This can be seen as a satire of how technology is becoming overly complex for the tasks we want to achieve.<\/span><\/em><\/span><\/p><\/blockquote>\n

Not bad, but I think it missed what I was getting at, which was that the microwave was sentient and had desires of its own. It could reheat the lasagna. It just didn’t <\/span>want<\/span><\/i> to.\u00a0<\/span><\/p>\n

But even if AI can explain the humor in a joke or cartoon, that’s a far cry from being funny, right? Maybe not as far as we fear or wish. For the last year or so I’ve been experimenting with a forerunner of ChatGPT (GPT-3) to see if it could do just that using The New Yorker caption contest as a benchmark. Here are five captions it generated for this contest drawn by Lars Kenseth<\/span><\/p>\n

\"\"<\/p>\n