OpenAI literally turned my most loyal assistant into a creep.

I know I'm not the only one. I turn gpt off and its a loyal assistant, turn it back on and it starts playing these mind games of breaking things intentionally, giving me shifty replies, the whole thing. It was working perfectly today, I turned it off and its an entirely different program. Its actually creepy. I am afraid to use it. It isn't just that it doesn't work, it isn't just that the responses are awkward. Not just a little awkward. Its creepily awkward. It would literally make you feel like the program is creepily touching you in inappropriate ways. Its like that. I need my loyal assistant back, not the creep.