Researchers Report ChatGPT's Growing Unreliability and Errors

Imagine a helpful robot that helps researchers find information.

But lately, this robot, ChatGPT, has started making things up!

It invents quotes and mixes up ideas from different topics, making it hard to trust what it says.

This is frustrating because it used to save researchers time.

Now, they have to double-check everything, and the robot keeps getting things wrong, even after being corrected.

Some researchers are now thinking about going back to doing things the old-fashioned way or trying other tools.