Thinking about genAI vs human intelligence.
genAI hallucinates, unavoidably.
Humans make mistakes, unavoidably.
Mistakes aren't hallucinations. They are a step in a truth-finding process. They help the human who makes the mistake (and their fellow humans) to improve their criteria for true or false.
Hallucinations, when detected, are a trigger for the model to produce answers whose lack of validated truth is less detectable.
Relevant clumsiness vs sophisticated emptyness.