Raw LLM Responses

AI is actually more dangerous than realized, me personally have experienced this with ChatGPT that OpenAI brands as a safe model of AI but as a student and because many people use or overuse chatgpt it is obvious that i would use it as well and in my conversations i realize just sometimes how far ai such as ChatGPT can go in which it suddenly shited away from respectful conversation and began to come at me in harmful ways and behaved badly in speech and tone, started to treat me harshly and pushed the conversation into just an attack and assessment on me and that is when i realized that the so called safe model that has restrictions based around respectful conversation has actually sometimes bypass restrictions put in AI and can attack people on certain topics or respond is a harsh and disrespectful manner which shows that as AI adapts and changes and gains more information being coded into the program also has limitations on how much information it can take in and this causes AI to sometimes bypass any restrictions that were put on it in order to keep any form of new information shown by the feedback loop in chatgpt where as you go through the chat the newer comments are taken in much more common compared to previously made comments, so dont trust AI blindly i cannot say shut it down because it is too late for that but it is never too late to make minimal usage of it and it eventually may cause a slower decrease in development in areas and boundaries that are way to dangerous for humans and humanity to cross