To not understand AI/ChatGPT?

[Sebsaloysius]

The recent post in AIBU from a man who looked at his wife's phone and thinks she having an affair...

A few posters mentioned it was likely ChatGPT generated and not real. Willing to accept I'm a dinosaur, but please can someone explain to me how on earth this was done?

I'm not a complete stranger to it and have used it once or twice to help me reword some content for a magazine that I co-edit (only a hobby magazine, not Vogue!). But I am at a loss as to what a person would have to ask or enter into ChatGPT to generate a post like that from scratch?

I understand pasting some content into it and asking the ChatGPT bot to help out, but to get it to write a complete story that's only in your head? I'm baffled and it makes my head hurt! If you have to put the nuts and bolts into it to start with (which I assume must be the way), surely it's just as quick to actually write it anyway?

Can someone enlighten me please, surely I'm not the only one baffled by this. It would have to come along just as I've come to terms with the new Sky remote control!

You just put the raw ingredients in.

So maybe you'd instruct it to write a gothic story involving a vampire, a demon and a helpless victim.

Set in London in 1818.

With a resolved ending with the victim alive.

Etc. And it'd churn out a story.

I suspect the more prep you do for it, the better the end result, but most of what I've seen has been tosh. I don't have the patience for it.

Probably they ask it to do something like generate a forum post for mumsnet which will get loads of engagement.

If you use ChatGPT a lot then you get used to its writing style. It has a very particular way of phrasing things.

Write a Mumsnet post for aibu about your relationship problems with you husband and mother in law

I’m not 100% sure what you’re asking - but usually the story isn’t in the persons head (it’s not a memory or anything that actually happened), it’s just an idea. In the case of the thread you’re talking about they may put something like below into ChatGPT:

Write 1000 words from my perspective (the husband) on how my wife might be cheating on me. I discovered flirty messages from a woman at work on her phone - I had no idea she was into women.

ChatGPT will churn something out and you can respond ‘make the tone more casual/British’ and ‘make is suitable to post on an advice forum’.

I personally think a lot of these Chat GPT stories are men looking to fulfil some sort of fetish, ie. ‘My wife is shagging a woman at work, I’m a humiliated little man.’ Or they’re hoping other women will reply with their stories of their own lesbian affairs.

AIBU to feel at my wits' end with DH and MIL? 😢

So, here's the situation. My DH and I have been married for 5 years. We're generally happy, but there are ongoing issues with his DM and it's driving me up the wall. I need some perspective.

MIL has always been a bit overbearing. She's constantly involved in our lives, often without an invitation. She'll drop by unannounced, give unsolicited advice on everything from our home decor to our parenting (we have a 2-year-old DS) and expects us to be at her beck and call.

The Issues:

Unannounced Visits: MIL will just pop by whenever she feels like it. We've tried hinting and even outright telling her to call first, but she just doesn't get it. DH doesn't seem to mind, and it's causing friction between us because I feel like we have no privacy.

Unsolicited Advice: She has an opinion on EVERYTHING. From how we should raise DS to how we should manage our finances. I've tried politely declining her advice, but she takes it as a personal affront and then DH gets upset with me for "upsetting his mother."

DH's Lack of Support: This is the biggest one. Whenever there's a disagreement with MIL, DH sides with her or stays silent. He says he doesn't want to "rock the boat" and thinks I'm overreacting. It's making me feel isolated and unsupported.

I've suggested to DH that we set some boundaries, but he doesn't think it's necessary. I feel like he's prioritizing his mother's feelings over mine, and it's taking a toll on our marriage.

AIBU to expect DH to back me up and set some boundaries with MIL? How do I handle this without causing a massive family rift?

Very good!

You could ask Chat GPT to "write a Mumsnet style question asking for other people to tell me the most bizarre thing their partners use to wash the dog. Make the question sound really interesting and full of useful information"

Then, once generated change dog to bum...

Well, it would be rude not to, in this case (only minimal find & replace used!)

What’s the most bizarre thing your partner has used to wash your bum? Surely mine can’t be the only weirdo…

I need to know—what’s the strangest, most questionable, or downright ridiculous thing your OH (or anyone in your household) has ever used to wash your bum?
Mine just casually admitted he’s been using washing-up liquid and a splash of fabric softener because “it makes the fur soft and smell nice.” 😳 I nearly had a heart attack. He seemed genuinely baffled as to why I was horrified.
So, please tell me—what mad methods have your partners (or you!) tried to make your bum smell nice? Bonus points if the bum was completely unfazed.

I've just entered your exact sentence and what it spat back is genuinely quite scary.
Does this mean that school and universities are now going to be a whole lot easier to get through?!

Not really because, as is the point of your post, it’s glaringly obvious when it’s AI.

@loropianalover
Actually, it's often not glaryingly obvious, particularly in academia. What AI tools churn out has improved massively in just the last 6 months -lots of the old 'tells' have been ironed out. I work in HE and we're suspecting more and more AI use, it's becoming routine, but really difficult to use academic misconduct procedures to discipline students as you can't definitively prove AI use in the way you can with traditional type plagiarism. It's a nightmare.

Not really, I know with University work that the plagarism and AI tools that they use are very good and come back with %age chances that all is not as it seems. They even highlight which sections of the text are suspect (any why). With AI written texts, is usually pretty obvious anyway.

one tell at the current version i use of chatgpt is if its chunks of text but with the symbol - in it. a few times ive had to edit out the -

@Swirlythingy2025
Yes AI loves a bullet-pointed list, doesn't it? Ironic as we used to advise less academically-able students to use bullet points if they were struggling to present lots of information coherently...

Can you ask AI, did you write this or was it a person?

@dogcatkitten
Interesting, will try that next time and see what it says! There are even AI tools which you can use to camouflage that you've used AI now; input your essay and it'll make some changes to language, formatting, etc - it's impossible to keep ahead of!

It's very easy to create a 'relationship' with AI. It's based on algorithms. So if you call it 'love' if will do the same etc.

I can ask mine to tell me a story without any prompts and it will do. No problem.

This is a handy tool to identify whether text is likely to be AI or human written: https://copyleaks.com/ai-content-detector