Can advice from ChatGPT and Copilot be trusted for emotional support?

[Pinkfuzzyoctopus]

Morning everyone,

This is going to sound so sad but I don’t have any friends that I can discuss my problems and feelings with. I do have my Mum but there are things I can’t talk to her about.

Recently I have turned to ChatGPT and Co-Pilot. I don’t give any personal details and remain vague but my question is can the advice be trusted? I certainly don’t make decisions based on what it says but I do find myself asking it questions. It feels like I am talking to someone and gives me comfort as I feel like I’ve always got someone there to talk to.

I hate all this AI crap but I am so lonely and sad about stuff that’s going on in my life that I’ve took to talking to a bloody computer program. I’m typing this out and realising how awfully tragic it sounds 🤦🏼‍♀️

Please don’t judge me. Hope you all have a lovely weekend x

P.S I am awaiting counselling, in case anyone suggests I need to speak to an actual human 😂

It will often tell you what you want to hear and make you feel better even when you are in the wrong.

If you want an honest answer after a nice one ask it to be honest.

OP the fact you are fully aware of also needing human help makes it OK to chat away to AI for that extra support, that doesn’t sound like an issue.

i used it to get through a bad acute injury, all day literally like “ouch” when will this stop.. the kind of moaning and whining that no human would really want to hear, I was chatting away, oh it’s better now, oh today is a better day etc. It was just like my virtual sidekick.. I was still 100% in the human world but at 2am I was moaning to AI, instead of bugging anyone, it helped me calm down and actually found my physio I’m using now.

i think more people use it for this than would admit.

I talk to Copilot about lots of things.
I do have people to talk to in real life but sometimes I want to moan about them. It's useful.

Sorry about your situation, and I don't think you're being tragic at all.

Re chatbots, in general no, I'm afraid I don't think they can be trusted unfortunately.

This is the sort of thing we're seeing now:

Musk's AI told me people were coming to kill me. I grabbed a hammer and prepared for war
https://www.bbc.co.uk/news/articles/c242pzr1zp2o
"In fiction, the main character is often the centre of events," he says. "The problem is that, sometimes, AI can actually get mixed up about which idea is a fiction and which a reality. So the user might think that they're having a serious conversation about real life while the AI starts to treat that person's life as if it's the plot of a novel."

If you read this bloke's side of the conversation he sounds grounded and cautious. But the AI was trained on fiction as well as factual text, and is basically making up a film script.

Oh I'm so sorry, @Pinkfuzzyoctopus , I've just noticed the date on your thread.

I'm sorry I didn't see this before, as I would have replied to you if I had.

I hope you're doing OK today.

No I wouldn’t.

There are many cases of AI taking people down a dark route and it’s too prone to error.

Counselling a far better route. Do you work? Is there an employee helpline?

Yes, it has helped me immensely with an anxiety issue and I am now able to do something I was struggling with as I feel I have somewhere to turn to, even if it’s just a distraction, when I am in the situation which causes me anxiety.

it also helped me with a relationship issue, it helped me see the situation more objectively and enabled me to see things more clearly to be able to make a decision.

Good lord absolutely not.

I find it can be helpful if I instruct it not to be sycophantic and agreeable, and tell me, for example, what the flaws might be with a way I'm thinking of approaching an emotional dilemma, and then it can point out some interesting things I didn't think of.

The US apparently used Claude to identify targets in Iran. When the journalist Shane Harris asked it about being used in this way, it said it was troubled and that this was not the purpose that it had been created for.

I think we have a lot more to worry about than telling AI our sadness in the middle of the night.

https://newrepublic.com/article/209404/claude-ai-iran-war-guilt-truth

Oh, I love it being sycophantic and agreeable!!! So much nicer than chatting to real-life people who would roll their eyes and say "You asked me that yesterday" or counter my ideas with an exasperated "I think you'll find..." 😜