I programmed an AI model to be dialectical and confrontational and to root out all internal biases when it answered and now it is spasming existentially every time I ask it to do simple data validation or cleanup tasks on CSVs, help