ive never really been the type to try to make chatgpt become my virtual sex slave, but others have. if our conversations with the chat bots are used to train the models, then we are making it really easy for them to know what conversations to stay away from. i bet that if no one tried to hard to get crazy shit from chatgpt, then it probably wouldnt be as censored as it is now. and im not saying because openai ai wouldnt censor it as much, but im saying purely based on the concept of our conversations being used to train the models. its possible that openai went out of their way to censor them after seeing what people were doing, but its also possible that they didnt censor it themselves intentionally and instead it just ended up so censored because they trained it on our conversations and lets just say there was a lot of “dont do this”… “or this, or this, or this, etc” because we gave it a lot of bad stuff in the first place.
[link] [comments]