Elon Musk's AI bot Grok has been calling out its master, accusing the X CEO of making multiple attempts to "tweak" its responses after Grok repeatedly called him out as a "top misinformation spreader."
If an AI isn’t in web search mode, it will just invent most likely answer, to the question of the source. Changes are very high that such sources don’t even exist.
That’s why you ask for the sources, so you can check them.
I think this kind of prompting is an important part of how to use it in any meaningful manner.
You can also input your own sources and ask it to only use that. For instance by uploading a pdf of a law and ask it to figure out how to do something totally legal and then let it show where in the law it says so.
You’ll obviously still need to check that the law actually says so and that it isn’t hallucinating.
If an AI isn’t in web search mode, it will just invent most likely answer, to the question of the source. Changes are very high that such sources don’t even exist.
That’s why you ask for the sources, so you can check them.
I think this kind of prompting is an important part of how to use it in any meaningful manner.
You can also input your own sources and ask it to only use that. For instance by uploading a pdf of a law and ask it to figure out how to do something totally legal and then let it show where in the law it says so. You’ll obviously still need to check that the law actually says so and that it isn’t hallucinating.