Ask a physicist why the moon does not fall out of the sky, and you will get a complete, honest, and verifiable answer. Depending on the current state of understanding, those facts may evolve. At one point in time, in some cultures, it was believed the Earth did not fall because it was held up by a turtle.
Ask a child how they broke a knickknack, and the storytelling machine turns on immediately.
“I don’t know. I was holding my ball when a big bat flew in and got into a fight with the dog. The dog jumped up to get the bat and broke the knickknack. It wasn’t me. It was the dog.”
Ask AI why you have toenail fungus, and you may get an answer that sits somewhere between an expert explanation and a child’s imaginative story. It will sound confident. It may even sound authoritative. But it is not guaranteed to be true.
AI is a remarkable tool for combining available information with the mechanics of storytelling. By design, it generates responses that feel complete, coherent, and satisfying, whether or not the underlying facts are fully settled.
Our local library frequently hosts historians who perform Chautauquas on historical figures. These presentations are a wonderful blend of historical fact, storytelling, and insight. Depending on the historian’s understanding of their subject and their ability to weave facts into characterization, the result may be spot on or slightly blurred.
And so it is with AI.
In all of these cases, facts can be slippery for many reasons. When they are combined with imagination, narrative, or generative systems, further investigation isn’t optional. It’s essential.
Even false or unverified stories are so much more compelling than data sheets.
Leave a Reply