Comment by 93po

I don't think it's as black and white as that. Giving messaging to an LLM and asking "How can I say this more clearly or more kindly?" gives valuable feedback to how you're communicating and how it could be done better, though obviously taking it with a grain of salt.

I think there is also value to affirmations and validation, even if it's done blindly by a robot. We have hurt feelings and want to feel understood. And when the source of those hurt feelings isn't immediately available to talk, it's a small tool to use for self-soothing behavior. Sometimes, or often times, these affirmations might be something you intrinsically already know and believe, and it helps to simply be reminded of them and worded in a different way.

To say "ChatGPT agrees with me and so I feel more confident that you're wrong as a result" is definitely the wrong approach here. Which is, to a small degree, what this person did. We did ultimately break up recently, and the reason being communication issues (and their unwillingness to even talk to me through conflict) is probably no surprise to you. But this outcome was very very likely regardless of LLM use.