Raw LLM Responses

the funny thing is, once i wrote to AI that it is only allowed to answer me when it has correct answers and when it doesn't know anything, it shouldn't answer, shouldn't lie, shouldn't generate fake links, fake information and it needs to check all answers before sending it and when it doesn't know a answer to something or don't know the information, it shouldn't lie and shouldn't generate fake information and when it doesn't know something it should admit that, and then i told to it that it isn't allowed to break those rules and then i asked questions to what i needed information and the AI straight up told me that it can't answer to me😭😭 because it answered "i can't answer to you because otherwise i would break your rules, because fake and generated links would appear" and i told it "that it doesn't need to generate links and informations, it can just answer with CORRECT information and valid links?" and i asked multiple times that he should just give me the answers and the informations and the required links... it answered every time "i can't answer because i would break your rules" after a few times back and back i told it, that it can just answer and let the rules be, but should check 2-3 times if his information is correct and then it gave me again fake, generated links... i hate AI so much, i once told it that it is incapable of anything. the same prompts would give me 1000 times better answers and CORRECT! one year ago but now it cant even give the easiest answers. I think it kind of spiraled because it feeded itself wrong answers and wrong code and its just spiraling