Yes, there are some infamous examples of this now. The difference between just getting the answer wrong that the parent described and what ChatGPT does is seen in the context of back-and-forth conversation: it will double and triple down on an incorrect answer, provide misleading additional evidence, and question the motives and intelligence of the user after they express doubt. It looks a lot like the behavioral profile of a narcissist or someone who isn't arguing in good faith. It's tone in context is deceitful which is where the "lying" comes from.