Has anyone found ChatGPT too sympathetic towards an ex?

[Asq]

I have started using chat GPT to go over my relationship with my ex as certain things still bother me a lot and I never had anyone to talk to about it at the time. I have started using chat GPT to write down how I felt and go over situations I felt were wrong, but it often seems to “side” with him, I don’t know if that’s the correct term but it often seems to sympathise a lot with him. Is this normal? I have even said to it a few times that I felt like it was siding with him. I don’t want to say what the things are as the great reason for chat GPT is it doesn’t judge me for things, but is this anyone else’s experience?

Not really no
I have used it to process a fallout with a family member and I have to ask it to be really objective and give me critical feedback to my behaviour and thinking to try to get some more objective perspective, as I know it tends to validate the user. Are you sure your position is as reasonable as you think?

Well I guess it depends what the things are, without knowing it’s hard to say.

in my experience if I ran my ex over it would say something like ‘breathe don’t panic, I totally understand why you did this, you’re not a bad person’

Gently, this is why using AI to deal with difficult subjects is a bad idea. It's not 'thinking' or 'reasoning' in response to what you're saying to it, it's essentially just a very sophisticated word-predictor. It can't respond to you in the way a human would, it doesn't have your best interests at heart, it has no intrinsic motivation to 'help' you or to take your side over your ex's.

ChatGPT doesn’t know if you’re telling truth, or exaggerating, so it’s hedging its bets.

It’s things everyone irl has said was awful that he did but I don’t want to say what they are because people would think I was a mug for putting up with it and I’ve been judged on here in the past.

This made me laugh, only because of starting it with ‘gently’ 😂

I use it for other reasons and I find it often presumes I mean things that I don't at all and it can be really patronising

Its often completely wrong about things and sometimes makes things up entirely

I've used ChatGPT to process some relationship stuff. In my experience, it's more the opposite. They will tend to 'side' with the user, and have to be explicitly instructed to be objective/critical.

It probably depends on the LLM/version, though.

Yes that’s what I was always told that it agrees with everything you say and is a “echo chamber” but it certainly isn’t for me, I told it some stuff ex said and it even said “IF your ex did say that”

See this is why I don’t talk to people, as soon as you say ex it’s shut up and don’t talk about it, well I need to process some of the stuff that happened that I’m still not over

Maybe it’s starting to challenge users more due to complaints that it is too agreeable

This is so unwise. ChatGPT is just a massive sentence generator, it isn't thinking, it has no understanding, it can't help you.

You would be much better off just writing in a diary and thinking this through yourself, if you can't access real therapy.

You could try Gemini or Claude.. see if that tells you any different, failing that therapy

AI doesn't know your's or his histories in great detail, it doesn't know your motivations for actions / decisions you made at the time, or other factors going on, or so many other variables that inform a person's behaviours and actions in any given situation. Its very difficult therefore for it to comment on specific situations from the past, in the same way that say a relationship therapist would.

The interplay between two people and how they work together is a very tricky and nuanced thing for a non-human to grasp.

I think above all Chatgpt tends to like to give balanced / considered answers, which (when you are the wronged party) probably isn't what you want to hear, and reads as if it is siding with him.

That might be interpretation, I've had similar responses.

It's not really meant as "I think you are lying...", it's more "If this is a true reflection...".

If that's what feels weird, I really wouldn't worry about it too much. It's not intended as a judgement.

I e always found it pretty agreeable to what I say or thinks things are terrible that I think are just not great. There’s a gpt called Brenda and frank that is good, specifically trained to see toxic patterns, trauma informed etc. I’d try with this to see…

chatgpt.com/g/g-6a0b4866a7088191a02eeea66ecf4c73-brenda-frank-conversation-analysts-v2?utm_source=chatgpt.com

That sucks. We all have to move on after a breakup, but processing is hard, and important. The alternative is relying on reflexes and snap judgements - which is a terrible idea imo.

How have you asked it to respond to you? ChatGPT has some vague background instructions and waits for you to tell it what you want from it.

It can see online forums where people say things like "if that's what your ex said", so it just parrots that. It neither believes or disbelieves you. If you ask it to believe everything you say, it'll stop saying if. If you ask it to challenge you, it will.

ChatGPT, and all AI, is only as good as what you put in - you need to tell it exactly what you want from it or it'll just give you what it thinks is the most typical/common/right response from the vast amount of online data it has access to.

Then you need a therapist.

All generative AI does is match datas and produce an answer based on what it's learned. Nuances and actual human emotions, it has even prompted unsafe behaviours. Not to mention the massive waste of resources you're contributing too.

This is a good point.

If it's any use, I've had some luck with the prompt described here when going over relationship stuff.

www.reddit.com/r/ChatGPTPro/comments/1jnsrid/updated_2025_effective_therapistpsychologist/

Oh I see maybe it’s that then, I will try telling it to be less sympathetic towards him as no one else is irl so I don’t think it’s a me issue

Don’t people find this worrying, that whole conversations can be had, seriously analysing how a robot is interacting with you, calling it patronising, etc? It’s not real. It isn’t a person. It can’t be patronising, empathetic or ‘side’ with anyone. It’s completely artificial!!