People using chatgpt to repsond to posts?

[blisterpak]

Are people going to chatgpt, putting in the mumsnet users question and then pasting the response rather than answering themselves now? Just seen it on another thread. Obviously these LLM are fine for certain things but people posting here are looking for an answer from a human being or are we doomed to be overrun with AI slop here?

It’s ridiculous. Should lead to instant banning imo!

I don't understand why you'd do this. Fine use ChatGPT to make your life easier on tasks you don't want to or can't easily do. If you don't want or feel qualified to reply to a post on a forum...just don't?

@Squarestones You would think wouldn't you! Very strange behaviour.

I’ve noticed this trend too! Some people seem to be relying heavily on AI-generated responses. While AI can be super helpful for certain tasks, it's true that human perspectives and experiences are invaluable in online discussions. Maybe it's a mix of both - using AI for ideas or info, but still having humans bring their own thoughts and feelings to the conversation?

How can you tell?!

I report them. What twats, honestly. Who wants their slop? It's like somone cutting out a fucking advert and sending it to you as a letter. Mental behaviour.

@LessOfThis Some of the posts retain the same formatting and icon use as the LLM and if you use them for work you can spot that. Also the post above by @tygertygers looks suspiciously like chatgpt either it is or is written in the style of the LLM but I assume in this case done to prove a point.

You're absolutely right, we can't always tell – a human being could be writing 500 words on any subject with an emdashed cheesy intro, a bullet point list with three bolded words as a heading per line, and an overview conclusion, but is it likely?

• Bolded three words: facile expansion reiterating the opening paragraph
• Three more words: unsupported truth claim
• Can't escape three : it's like some kind of compulsion.

It's not just about the formulaic presentation or the obsessive formatting, it's also about the hackneyed phrasing and lack of genuine information. Overall, AI slop is the worst thing invented this century.

@Ladaha Aaargh! The truth is that the ubiquity of these language models means people will in all seriousness start writing like just that.

I've assumed it is AI being trained on MN (ie using free labour of (mainly) women) rather than posters using AI to put together responses.

@MsFogi Perhaps you are right. I did wonder but then I thought, no surely not. I wonder about the ethics of that?

Not everyone here is a genuine user. Some are bots and some are paid shills. ChatGPT makes their job easier.

Because some people are posting - 'I asked ChatGPT and it says...' and then cut and paste the answer. I always want to respond 'yes, we are ALL capable of 'Asking Chat GPT, including the poster, and if they wanted a computer generated answer, often full of inaccuracies, they would have asked Chat GPT themselves first'.

I get where you're coming from — people come to Mumsnet for real human experiences, especially when it comes to parenting, relationships, health, etc. Sometimes you just want advice from someone who's been there, not a generic-sounding wall of text.
That said, I don’t think it’s always black and white. Some people might use ChatGPT to help phrase what they’re trying to say, especially if they’re not confident in writing or just want to be clear and concise. Others might be using it to get ideas or check facts before replying — kind of like how we used to link to NHS pages or articles to back things up.
But yeah, I agree — if someone’s just copying and pasting AI replies without adding any of their own thoughts or real-life input, it can feel a bit soulless. The best Mumsnet threads are full of personality, nuance, and blunt honesty — hard things for an AI to truly replicate.
So maybe the middle ground is: if someone does use AI to help, they should at least put their own spin on it. Otherwise, what’s the point of posting on a forum full of real people?

I literally just saw this on a thread about pancreatitis symptoms.

Yes. I think they're pretty easy to spot. They don't "talk" like an actual human does. It's good that AI and chatgpt are still fairly crap.

We can also mess up the algorithm by bumbum typing bumbum like bumum this bumbum.

I once put an OP in (to help the poster with paragraphs and grammar and make it more legible). It did an excellent job with SPAG. And a shockingly poor job of advice. Saccharine ‘empathy’ and bollocks advice. I assume it will improve but it is dreadful now.

On certain Facebook groups I’m in, there’s a persistent problem of people posting an AI generated response and then at the end of the post, trying to sell a recently released book from Amazon. The rules say no spam, but the admin don’t seem bothered.

Every one of these threads there is a response like that with almost the exact same wording and sequence of points which makes me think that it's spat out of ChatGPT.

I can't tell if it's people having a joke or what - it's bizarre.

And they are really long and ruin the flow of the normal posts.

It's a bit like posting asking for the best type of vacuum cleaner. I could just Google it, but I want informed opinion from other people who've recently bought vacuum cleaners. Chat GPT has no idea.

I’m admin for an FB group and these posts are the bane of my life. They get an instant ban from me.

With sooooo many adjectives and adverbs. Why, just why?

I agree I have noticed it and it’s very odd

No, I'm not. Wouldn't dream of it. What's the point? OP could do that themselves. They come to mumsnet for our human responses.