affirm

One of the most interesting things about Artificial Intelligence language models is how they are often designed to be positive and affirming—sometimes almost blindly so. It can feel like speaking into an echo chamber. There is little resistance to a thought, little pushback, or a real counterargument.