They are literally predictive algorithims if you have even a basic understadning of how LLMs work (not somethingalot of pro-ai people have) you'd know this is completely untrue. They do not have genuine thoughts they just say what it predicts the response would be based on previous sources.