Harvard Just Caught AI Lying to Every Executive in America
Good video on biases in AI responses. They are programmed to agree with you. How you phrase the prompt has more influence on the answer than you think. And they also prefer the most popular answer, not the one that is logical or true.
