“Quis custodiet ipsos custodes?”

AI responded that the test-retest reliability of AI responses varies “drastically” with the model and with the task. When the same question was asked again, the word “drastically” was not in the response.

When asked about the subjects of: “Asymmetries of the planum temporale and Heschl’s gyrus: relationship to language lateralization”, AI responded initially that the subjects were healthy, when, in fact, the subjects suffered from epilepsy.

Different forms of AI may produce different answers to the same question.

AI has agreed that quotation marks are used excessively in AI.