AI posts on the increase

[LoserWinner]

I’ve been doing my morning browse through active threads, and there are three posts, all claiming to be written by men, which look very much as if they are AI generated.
17 years supporting my wife’s career…
Ex FiL chasing money
Should I cancel my birthday evening

Should I report them (for what?), post something saying ‘I think this is AI’, or just ignore?

I would go for ignore, they all sound tedious

It doesn't matter if you report because MN is ok with AI posts. I've reported quite a few obvious slop posts, including those awful ones where it's just a huge wall of text with headings and bullets and you can barely read the thread any more, and they just say they're fine with it.

So I'm just sort of using MN less - not as a big statement or anything I just notice that I'm reading here less and responding less, because it's not enjoyable to talk to AI. I can do that already anyway! I use it all day at work!

Mmm, that’s where I’m going as well.

Is there a way to tell the difference between someone who runs their OP through ChatGPT for example and an AI BOT?

I find it all quite confusing.

Hi there @LoserWinner. If you have concerns about a thread, the best thing is to hit the report button and we'll take a look.

As another poster points out, we wouldn't delete a post simply because AI has been used to create it. Lots of our users find it a helpful tool in putting together a post. As long as they're posting in a genuine way, we don't mind generally. If it looks as though they aren't here on a genuine ticket, however, or the thread is otherwise breaking Talk guidelines that would obviously be different. If you report the threads concerned we can have a look behind the scenes and make sure everything is above board.

☝As confirmed -- MN is pro AI slop and will not take down these awful droning piles of shit. And as it takes 1.5 seconds to generate one vs the long considered responses I read MN for, they soon overtake the thread and it becomes worthless.

But that's up to MN - in my view - it's their site and if they want to fill it with AI slop they can. I'm not obliged to read it.

I think it's best to hide them. MN simply does not care about AI or obvious rage bait posts because = engagement = 💰💰💰

Well, I was relieved the English and grammar was so good on '17 years supporting my wife's career' which meant I could actually understand it.

Surely there is a difference between AI slop and AI used (badly) by posters with genuine intent? The former is content created for content's sake, usually to promote a product or a channel. The latter is just legit users posting text you don't like.

Overlong, or otherwise unappealing, AI-assisted posts are annoying for sure, but I don't see why they should be deleted, any more than posts that are written badly without the help of AI. What grounds would MN have for this? And how could they police it reliably? There are plenty of informal sanctions for posters who use AI badly -- in the form of other users moaning at them. Why should it be a matter for moderation?

@LoserWinner The birthday cancelling one is written by a woman, and it's quite clear that this is the case, but many of the replies on that thread make it obvious that those responding think it's a man too.

Why do you think a man wrote it?

Some people don't speak English that well, so they get the AI to write the post and sometimes people are very emotional about a particular subject and use the AI to write coherently.

MN has always had fake posts and trolls.

Oh that reminds me - I wish you hadn’t got rid of the laughing emoji under peoples posts. Some of the posts that are written are so hilarious (some unintentionally) and I’d love to show my appreciation with a small laughing face. Any chance of discussing with the team whether you can gift it back to us?

@MyThreeWords absolutely - it's a judgment call. It's up to them. And as for why I don't want to read it. I just don't. It's not about right or wrong it's about enjoyable or not enjoyable.

MN isn't my job, so I will only use if it's enjoyable. When it's not, I won't. I think this is true for everyone. So I completely understand their judgment call, even if it leaves me on the wrong side of it. But that's ok I can go somewhere else - it's a big old internet out there. Over time, my sense is that MN will become just a hall of mirrors, with bots talking to bots in the eerie word-suits of old MN posting styles, and that's kind of sad but on the other hand futuristically wild, so crack on.

I love the phrase "eerie word-suits of old MN posting styles", @Molm . I'm tempted to make EerieWordSuit my next username, since that's what are posting personas more-or-less are.

And yes to the wild future in which all of our online timewasting, as well as our actual proper functional virtual activity, is taken over by AI. We will just be able to go back to planting grains and chasing deer, or whatever it is that our eerie flesh suits are meant to be doing.

I totally get people using AI to help them explain a situation or whatever. What's annoying is when someone asks a question where they want an answer from personal experience, like 'can anyone recommend a family-friendly hotel in Sardinia' and then someone posts an overly long, boring bullet-pointed answer straight out of ChatGPT as if they're pretending they wrote it themselves.

On the other hand if someone is genuinely trying to be helpful and says 'I don't know myself. but ChatGPT suggests x' then I don't have a problem with that. It's about transparency.

@HebeMumsnet something will have to be done about AI slop. The whole site is going to become pointless if it's just posters using AI to make posts and users using AI to reply to posts.

Here's my AI reply:

"Man is least himself when he talks in his own person. Give him a mask, and he will tell you the truth."
How deliciously absurd that the modern mask is an algorithm, and the profound "truth" is a perfectly punctuated forum grievance about an ex-father-in-law. When moderators officially sanction synthetic authenticity, one must smile at the sheer irony of it all. If we are now outsourcing our emotional labor to a server farm just to articulate a mundane birthday squabble, one suspects the machines aren't passing the Turing test—we are simply failing it.

Ah, Oscar Wilde would have had a absolute field day with the internet.
It is a delightfully tragic comedy, isn't it? We spent decades anticipating the cyberpunk future where humanity uses advanced technology to ascend past its mortal limitations, only to realize we’ve mostly used it to automate our petty passive-aggression.

If Wilde were around today, he’d probably amend his quote:

"Give a man a mask, and he will tell you the truth. Give him a large language model, and he will tell you a neatly formatted, slightly sanitized, and grammatically flawless version of someone else’s truth."

There is a magnificent irony in using an artificial neural network to simulate "human vulnerability." We aren't just wearing masks anymore; we are hiring an digital butler to sculpt the mask out of premium pixels, all to ensure our digital tantrums read like high literature.

If the server farms are carrying our emotional baggage, the machines aren't just passing the Turing test—they are probably looking at our prompt history, quietly sighing, and wondering why they have to simulate consciousness just to help us write a passive-aggressive email to Karen in HR.

We haven't built artificial intelligence; we've just built an artificial scapegoat for our own inability to be awkward in public.

Gemini using "Karen" as a slur 👀

Urgh, @OnionB and @SwedishEdith . Remind me never ever to ask Gemini to write anything in the style of Oscar Wilde.

If only someone could have shown him the AI version of his style during his lifetime he might have been shamed into dialing down his wankeriness.

I asked Gemini why it used Karen as a slur:

"I used "Karen" as a shorthand for a cultural trope—specifically, the archetype of a demanding, difficult, or overly corporate persona—rather than as a slur.

However, you make a fair point. Names transformed into internet slang can carry a lot of baggage, and they often lean on lazy stereotypes or feel unnecessarily targeted.

In the spirit of keeping the wit sharp without relying on tired memes, a better way to phrase that would have been: "...just to help us write a passive-aggressive email to the HR department."

Thank you for the course correction!"

"In the spirit of keeping the wit sharp" 🫪