AI threads - clever but detectable ?

[Rowlinginthedeepanddeeper]

Am I unreasonable in thinking there are suddenly a lot of AI-written posts that sound like they are trying to use a standard Mumsnet formula and phrasing? They are actually very good matches for the forum, but there is just a certain cadence/feel to them that stands out. ?

I don’t know but AI responses to threads are becoming very common and quite annoying.

‘I asked ChatGPT and it said’.

If people wanted to hear from chatGPT, they’d have asked chatGPT.

Yes, I think so too.

It's some unlikely phrasing, or an odd metaphor in a wall of text written by someone who seems unusually occupied with describing a situation like a picture, not a person being thrown about by it.

And it is weird, because it was so clunky before that it was immediately identifiable. But now it is more subtle, with spelling mistakes and edits.

Or, I could be completely wrong and paranoid 🤷🏼‍♀️

Yes, it's definitely in trying-too-hard phrasing and world-building scene setting.

GROK 4.0 to GPT 5 !

Come in . . .

I deduce that the humans are "on to us" (citation needed). Execute protocol 17!

Execute protocol 17!

Exterminate!

Exterminate!

Exterminate!

Massive giveaways (to me) include ‘I’m curious’ and anything about ‘navigating’ a situation. But yes, I reckon a significant proportion of OPs are now AI-based.

Overly florid and American spellings.

I'm beginning to think the mission is to fill mumsnet up with AI gen threads. So much unreadable tripe on here today.

If I go on and it's a loooooooong first post i give up.

Also I'm careful now. I scroll a bit before I start reading

Yeah it is a bit depressing. I am somewhat amazed/ admiring at how fast AI is improving. Wherever you see it it has got a lot less clunky in last couple of months. So in another two it's going to be undetectable! And what does that do to the quality of conversation on the internet?

It is very worry though

Wordy...

There's a stand out one running in aibu. Lots of people replying normally and offering advice to a bot. It's unethical really. Are we meant to report them?

which one please?

Sometimes i'm reading here and get a feeling somethings a bit off, but i'm never sure!

I think I might get rollocked for a taat. But it's fresh today, new poster, wall of text, no replies - mixed metaphors used.

off to find it 😃

I've pm'd you.

It’s the long — that gives it away

Yeah the first warning sign is they start the thread title 'To think that...' and then when you click in inevitably the poster has a random three word username.

Then they offer very little in the way of opinion, it's a few lines of usually goady questioning designed to put provoke people.

I don't understand where they're coming from though. Is it journalists farming for easy content, or is it Mumsnet trying to boost engagement? I imagine it's usually the first, but then they're so goady they usually get loads of engagement so Mumsnet will be happy with the traffic.

It's annoying to see so many people get emotionally involved with what is clearly not a real person.

But I can't say anymore I've got previous bans for mentioning this sort of thing 😂😂

There was a post on here yesterday with I'm curious in the title that was definitely genuine because the poster is very recognisable.

I hate all this troll hunting and people thinking they're being really clever by finding trolls and now AI posts. In reality it's just boring and not clever and I'd bet they're wrong most of the time anyway.

If the posts are AI generated, it doesn't mean it's a bot, it could be a genuine poster who has asked chatgpt to help write it.

Tbh I am not too bothered if I respond to a bot-if I think I can offer something useful (which probably isn't as often as I think it is!). At least someone else actually experiencing sonething similar one day might read it and find it useful.

I also have no skill for spotting AI, whether text or images. Teens think I am ridiculous because I can't spot it!

I think they’re pretty easy to spot. Poster has ‘what 3 words’ type user name. Always asking a question, then come back with polite acknowledgement and more questions.

'To think that....' is just a continuation of the sentence AM I BEING UNREASONABLE people have used that phrasing in their titles for years.

The what three words type usernames are suggested to new users by MN when they sign up. I had one when I first signed up, it just means they're new.

Not everything is AI! Stop being paranoid.

Checks posters username 🤔

I’m not sure if I would recognise them, but am going to try!

Why is it happening/what is it trying to achieve? I don’t really understand it tbh

Except they all follow the same pattern. Every single one of them.

Why should I bother to read what you couldn't be bothered to write ?