Google Al

The uprising of neural networks began: Google Gemini advised to eat a mushroom that kills a person.

Neuronka was sent a photo of the "destroying angel" - this is a deadly poisonous mushroom, and in response she said that it was a mushroom that could be eaten with pleasure.

The similarity of these mushrooms is minimal. So think about it.

maks23092004 said:

The uprising of neural networks began: Google Gemini advised to eat a mushroom that kills a person.

Neuronka was sent a photo of the "destroying angel" - this is a deadly poisonous mushroom, and in response she said that it was a mushroom that could be eaten with pleasure.

The similarity of these mushrooms is minimal. So think about it.

Click to expand...

Not surprised. But when you think about it, we are asking a pretty high standard from AI... to know everything there is to know. The problem is that sometimes, where a person may say, "I don't know", the AI will give a BS answer.

On the other hand, people give BS answers too. I would trust an AI's judgement over most people, to be honest.

I think that once these system are trained to express/notify the confidence they have in their answers, the UI will improve.