I understand how my comment was unclear, but I was attempting to underscore the fact that it cannot determine the difference. That’s why I included the whole statistically plausible bit. My point is that AI as it currently functions is fundamentally flawed for several use cases because it cannot operate as it “should”. It just says things with no ability to determine the veracity.

The first portion of my comment was addressing the suggestion that there was a “reason” to lie. My point is there is no good justification to providing factually incorrect answers, and currently there is no way to stop AI from doing so. Hope that clears things up.