Gemini Pro fails the famous 6 sisters vs. 1 sister test

It's broken on so many levels at the moment that it's unreal. Try asking Bard what model is it based on right now and you'll get all sorts of imaginable answers from PaLM2 to LaMDA to a "mixture of models" alongside the Gemini Pro, and sometimes it reverts back to "I can't really tell because the model is undisclosed, blah...". And later on in the chat you wonder why is its code examples so bad, and you ask it again about its model, and it's been reverted back to PaLM2/LaMDA again. sigh

Does not look promising at the moment, especially when you can't tell whether it's hallucinating, if it's even really Gemini Pro or not, because it can literally "change its mind" on "what it is" mid-session.

Not like this, Google AI.

Each of those brothers have Kaley as sister because they are her brothers. Now assuming they all have the same 2 sisters, then any sister that Kaley has would need to be one of them, so Kaley would have one sister.

But what if each of the brothers have a different mother or father than the the rest of the group does? Kaley could then have any number of sisters none of which would be shared with her brothers. So what is the correct answer?

So this is tricky even for real intelligence. The problem with the AI seems to be that it cannot explain its incorrect reasoning (because it doesn't do reasoning, only statistical matching) and thus cannot realize that it has the wrong answer.