Bingo
Bingo
It’s only going to get worse because we’re further away from media literacy than ever before.
You're viewing a single thread.
We're basically talking about truthiness.
People fact checked a crazy uncle if the things their uncle was saying weren't truthy. But, if a friend's medical advice sounded truthy, they simply accepted it and didn't bother to check.
ChatGPT and other LLMs are designed to sound truthy. If you asked "What's the biggest planet in the solar system" they could respond "I think that the biggest planet is Jupiter. My information is that Jupiter has a mass of X and a volume of Y. The next biggest planet should be Saturn, which has a mass of Z and a volume of Q. But, you should verify this by consulting these sources..."
Instead, these LLMs are designed so that the answers they provide are incredibly confident. When a user replies that they're wrong, they change their answer and generate a new one that is just as confident. That's the kind of answer that is going to seem truthy.
That sounds correct enough that I won’t bother checking the link