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?

Oh that makes sense as I think it said something similar to me when I asked why it kept siding with my ex. And yes the constant IF that did happen, IF that was said. I will try the other one I hadn’t heard of that one thanks.

Quite difficult when you have children with the person which serves as a constant reminder every day.

ChatGPT is generating text based on patterns it learned from training data. You are treating like it has a brain, and is making judgement calls on what you are writing. It doesn't understand what you are writing on a human level, it is creating sentences that seem appropriate in the circumstance. So no, it isn't bias towards him, it is using patterns it has gathered as a basis to create words. If you want to be more sympathetic/ use a specific tone then tell it, and it will create text based on those preferences.

Ok what is abundantly clear is that you are very unhappy. Honestly, please invest in a therapist. ChatGPT is not telling you what you want to hear so you assume that it’s being “too sympathetic”.

We are maybe telling you things you don’t want to hear like ChatGPT and again you are on the defensive and think we are wrong

That's even more reason to work on moving forward with your life and not dwelling on old grievances until you get the "right" answer which you want to hear.

He's not just your ex who you want to hear slated by friends/robots, he's now someone you have to work on coparenting together as best you can. I don't see the benefit of going over things which upset you at the time and trying to decide "who was right". It doesn't matter, you broke up, noone has to "win" the argument now.

And please don't put private information about your children into AI, even if it's too late to stop putting your own secrets in.

I don’t want him slated. I just don’t want it to try to justify cheating and other stuff. I don’t want it to slate him.

It knows I have children, that doesn’t bother me.

You’re looking for a human response from a machine. It’s not sympathetic to him, because AI doesn’t have feelings - it’s literally creating patterns based on what you tell it. It’s maybe not the tool to use if you can’t separate out an AI machine feeding you stuff from an emotion based response you might get from a person. In short, it’s not real and therefore can’t sympathise, empathise or judge, or give any other emotion based response.

Then write a prompt before you start the conversation telling it not to do this and saying what kind of responses you want it to give. You can shape the response it will give you through giving it a prompt. Your prompt can both include what you want in a response but also what you don't want. If it doesn't do what you want, then you need to work on your prompts so that it can generate the text you are looking for.

Yes I will do that.

Just to be clear here... you're talking to a machine about your feelings but you're not happy with the advice it's giving you, so you're now on the internet asking strangers for advice on how to get the machine to give you the advice you want?

Respectfully, I think it's time to snap the fuck out of this. Go outside, enjoy the sunshine, give your kids a cuddle, spend time with people you love doing things you love. Make a happy life for yourself and your children.

It almost feels sorry for him, this man was abusive to me (won’t be going into details) if you wouldn’t be questioning why it seems to feel sorry for him then I don’t know 🤷‍♀️

I'm more worried about you and why you are doing this to yourself. I don't think trying to understand AI so you can manipulate the answers it gives you is in any way helpful to you.

No need to worry about me, I will tell it to be less sympathetic it’s helped a lot and I think I did myself no favours telling it to begin with

I disagree. I find it really helpful. It structures my thoughts, reframes my perspective and challenges unhelpful thinking as well as offering simple practical tips. It also remembers what I said at previous ‘sessions’ which makes it better than at least one human counsellor I have had.

It’s brilliant at helping me in crisis situations. It says all the things a good therapist would, without me fearing judgement and without crap advice and I don’t need to pay it or wait for an appointment. It’s there when I need it, in the moment.

I’ve tried a lot of therapists, some have been actively damaging. Chat gpt has been better than all but two, and only one of those two would I chose in preference to chat gpt.

I suspect it will reduce the demand on human therapists, leaving therapy for those with more complex or long term needs.

It is absolutely doing that. You told it to when you said that.

No but then I only ever use it to workshop song writing and I often wonder if its blowing smoke up my arse regarding how 'brilliant' they are. I've thought about telling it it doesn't need to do that but who am i kidding, I like the praise really. Maybe you could just tell it to stop taking his side!

Yeah my mistake, I’ve realised since I wrote it down that I’ve obviously told it I liked it 🤦🏻‍♀️

It’s a TEXT PREDICTION machine. It has zero intelligence. It doesn’t think.

It says all the things a good therapist would

It isn't just about the words, which are utterly meaningless from a machine. Therapy is a human-human dialogue. It's just empty drivel from a machine.

This

In my experience ChatGPT and other AI tools will ALWAYS initially give a fuzzy sort of comment somewhere in its answer about someone however nasty the person you've told it about was.

E.g. "OK, he's a serial axe murderer but he loved his children so that's good"

Simply because one quote amongst a gazillion awful ones about him chopping his victims into bits was found where one of the axe murderer's kids said "but he was such a good Dad"

Don't let some nobbers on here stop you posting or interacting with humans. I only tend to get frustrated if kids are being involved in their parent's stupid choices of partner and are caught up in it but we all make mistakes and at least he is an ex now so you did what was needed to be done.

No I definitely won’t post about it on here again people on here have been vile to me at least chat gpt doesn’t judge

As someone who works in AI, mostly for codegen and SRE use cases, I think it’s frankly insane that people are using it for the use cases described here. It’s a fancy text completion engine, nothing more. It doesn’t think, it doesn’t understand, it doesn’t have empathy. It just takes the conversation you have had so far and prints out the most likely next response based on its training data. Well, until you exceed its context window and it starts to forget earlier parts of the conversation.

Sam Altman and his ilk have really pulled a number on people haven’t they?