'Calling out' AI -really necessary?

[WarmthAndDepth]

Can we not be so quick to 'call out' posts which appear to be AI generated, reporting them and more often than not getting them deleted?

I am no fan of AI, but understand that many people now use it for many things, apparently also to create and respond to posts on MN.

It may be tedious and impersonal, but if someone feels more confident constructing an argument or organising their thoughts in an opening post using ChatGPT, is that really so bad (providing the poster is a person and not a bot)?

It seems contradictory that a woman asking for advice about something deeply personal in mangled syntax and a smattering of spelling mistakes, omitting to use paragraphs or basic punctuation basically gets eye-rolled into oblivion with trite remarks like "Couldn't understand a word of that!" while posts which appear to display the hallmarks of AI (a certain slickness coupled with grammatical structures straight out of a Govian SPaG paper) get reported.

To be clear, I think casual use of AI is environmentally irresponsible and would much rather it wasn't 'a thing' but calling out its use on a site like this where a cohort of posters repeatedly shit on others whose erudition isn't deemed up to scratch isn't constructive or inclusive.

@MNHQ On threads where there have been 'accusations' of AI generated posts, and MN closes the thread 'to take a look'; what actually happens?

I’m sure there have been other threads about this.
It’s really no different to anything else - posters shouldn’t be ‘calling out’ ie troll hunting on threads. If someone thinks a post isn’t genuine then it should be reported. IME MNHQ does want to know about posts which really don’t seem human, they can then take a look and decide what they want to do.

The thing with AI is that it’s essentially making up a story. So yes, so threads are generally fake, even if the poster had the right intentions.

And use of AI is bloody lazy and is starting to come back to bite people now.

We need to be careful that people don’t lose the art of actual communication because they no longer have to as ChatGPT can do it for them. It’s bad enough that people no longer communicate face to face as they can do it by text or via social media, now we have AI they don’t need to do that any more, and it’s definitely not something that should be encouraged.

I know several people now who, if it’s obvious a job applicant’s cover letter is written using AI they are automatically rejected, and that’s how it should be.

Yes, I think I have seen a thread complaining about opening posts by AI bots, suspected of training LLMs, which is fair enough.

But when anyone who is a bit verbose is 'called out' (thinking about a particular thread this morning, poor guy) and reported, it seems excessive.

I was told my response to a thread must be AI not long ago -it's ridiculous.

A lot of the posts that I think are AI come across as spam - ‘Hey OP, brief reference to thread topic and positive comment! Have you tried this product?’

They’re irritating so yes, I usually report them.

@Never2many I agree about covering letters etc.

But is it really 'making up a story' if someone goes through the trouble of inputting the particulars of their relationship or parenting issue into an AI generator and posts the 'polished' version on MN?

I generally think that most AI posts on MN posted by actual people tend to be opinion pieces where a poster may use AI to help them structure an 'effective' argument, such as threads about VAT on private schools, immigration etc.

On an online social chat forum, where there are no consequences concerning misrepresentation in employment etc, should we effectively discriminate against members who feel their writing could use a helping hand?

I think it should be part of netiquette (there's a word I've not seen in a decade) on discussion forums to not use AI to generate your posts. People are expecting to interact with other humans and other social media is being ruined by junk content (see facebook and twitter), and I'd like to see MN as an AI-free zone.

It should also be netiquette for posters not to be a dick about bad grammar in a way that might push people towards AI.

@EmpressaurusKitty Yes, and that is covered in the Talk Guidelines too.

I guess I am talking more about threads where a person's posting style comes under question as opposed to the content. This really isn't a TAAT, but the thread I saw this morning was reported as AI whereas it was evident to me and other posters that the OP's idiosyncratic posting style arose from cultural norms and alluded to ND. Surely we need to give people the benefit of the doubt?

I did this last night, oops.

You're right, I shouldn't troll hunt on a thread. But what I commented on was a huge block of text that contributed zero to the conversation. It wasn't someone using AI to structure their argument, there was no argument, it was just a pointless lengthy statement of some of the aspects of a political situation 'on the one hand and on the other hand'. It was like someone fly tipping on the thread. God knows how teachers are coping with this shite.

@noblegiraffe "It should also be netiquette for posters not to be a dick about bad grammar in a way that might push people towards AI."

Yes, absolutely. I can't stand marginalisation and snobbery around language. I'm an immigrant and, like you, work in education so this is a real bugbear for me.

If you really want to see AI posts, try the Politics threads. It's perfect for training a real person in how to spot a bot. AI posts are will not be good for MN in the long term as it will make us question all posters.

There's a thread here on a similar theme, MNHQs response is clear they'll treat 'calling out' of ai-formatted responses as troll hunting

www.mumsnet.com/talk/site_stuff/5272369-ai-hunting-on-the-boards

And otoh they do delete and ban ai-bot posts.
They could do with some explicit guidelines on legitimate vs inappropriate use of AI tools but really it's quite simple, they just need dealing with like any other issues - report rather than police on-thread.

It really is the height of laziness to use AI to post on a forum.

@ErrolTheDragon Thank you for the link, that's a useful reference.

I chuckled at this quote from the OP @SoManyTeeth
"I understand the urge for posters to demonstrate that they haven't been "taken in"," -this is so often what it seems to boil down to, isn't it?

Nope. That post was blatantly written by AI.

Another giveaway is often that an OP writes an endless post and then disappears completely. So it’s just AI generating a discussion.

@DailyCake That's interesting. I hang out in Politics quite a bit but have largely attributed AI-esque posting styles to people's tendency to emulate the styles of writing of the literature or print -or online media they consume. That, and generally being current affairs nerds. I'll take a closer look...

Is this about the supposed Chinese American poster with the autistic son that they didn’t want to send to their dream highschool?

That OP was shockingly fake/long/over-involved and by the time PPs have got to the bottom of thread and realised, they’ve wasted time and energy. I think we need to know that OPs are made in good faith and if calling out blatantly AI generated (virtually phishing) posts helps to maintain our confidence that we are interacting with genuine humans, then it’s fine. On that particular thread one PP actually fed the whole OP into AI to get a very nice and succinct summary… which is how AI/ChatGTP should be used and would brook no pushback from me, personally.

@Never2many and @CautiousLurker01 That's so weird. I found it read -stylistically- very similarly to correspondence I'm currently processing for secondary school admissions from a small group of parents of the same ethnicity as the OP and advising on how to make their appeals more succinct and to the point. For this reason, that post rang totally true to me, right down to the longwindedness and some of the apparently irrelevant inclusions of detail.

Maybe they are using ChatGTP to compose their correspondence, too?

@CautiousLurker01 ChatGPT generally doesn't produce 'walls of text' though: I tend to spot it in my students' work as it appears grammatically flawless, quite smug and almost too 'smooth'; like an unholy conglomerate of George Clooney and the Cheshire Cat.

I don’t know - I just asked chatGTP to write a post for Mumsnet with a simplified version of the scenario of that post and it was pretty long and reused many of the same phrases. I’m totally convinced that post was AI produced.

I'm a bit conflicted about this. The issue for me isn't the bots, it's that posters - who maybe vulnerable themselves - interact with them in good faith.

There has always been a problem with trolls on MN, whether it's people having a laugh, begging threads or activists looking for screenshots. And equally there have always been gullible people who respond - myself included on occasions.

I don't have an issue with posters using AI to describe an issue they want advice about, but I do worry about the impact some of the fantasists and bots have on some people's mental health. And so I think helping posters who may have invested in a thread to understand that it is AI generated can be important.

Agree

@CarefulN0w I'm sure I too have engaged in good faith with 'bad faith' AI posts on many occasions (one of which I'm pretty sure is still going strong this morning). I don't really mind. What I find irritating is that a thread which has generated lots of thoughtful and insightful responses might go 'poof' on account of the OP being accused of AI-ing and all of the collective wisdom is lost.

Agree with this - threads aren’t owned by the OP are they? When PPs have engaged with the subject matter and each other and produced some really thoughtful exchanges it is upsetting to see those gone. I think it should be enough for MN to just put in a post flagging that there are some concerns of the authenticity of OP, if they agree it seems AI generated, but allow PPs to continue their sub-conversations?

I understand why people would want to use AI to write their posts but it should be flagged as such, then other posters can choose to respond or not. Then either nobody will respond and the problem will go away by itself, or it will turn out people don’t mind and we can all carry on.

I personally hate AI posts and don’t reply to them, I sometimes report if it’s obvious or just leave it if I’m not sure. I’ve almost stopped using Facebook because of the AI junk posts, it would be a shame if MN went the same way.