The word 'lie' is problematic here because it anthropomorphizes AI behavior.
A lie implies:
Intentional deception
Knowledge of the truth
A deliberate choice to mislead
When AI systems confabulate or fabricate information, they're not 'lying' in this sense. They don't have beliefs about what's true or false, nor do they have intentions to deceive. They're generating statistically plausible text based on patterns in their training data.